F A C U L T Y O F S C I E N C E

U N I V E R S I T Y O F C O P E N H A G E N

PhD thesis

Nina Molin Høyland-Kroghsbo

Interspecies quorum sensing as a stress-anticipation signal in E. coli

Academic advisor: Associate Professor Sine Lo Svenningsen

This PhD thesis has been submitted to the PhD School of The Faculty of Science,

University of Copenhagen, Denmark, 27 October 2014

Bacteriophage λ, image credit: Institute for Molecular Virology, University of Wisconsin, Madison

Bacteriophage λ, image credit: Institute for Molecular Virology, University of Wisconsin, Madison

2

PhD thesis title: Interspecies quorum sensing as a stress-anticipation signal in E. coli Author: Nina Molin Høyland-Kroghsbo, M.Sc. Department: Section for Biomolecular Sciences

Department of Biology Faculty of Science University of Copenhagen Copenhagen Biocenter Ole Maaløes Vej 5 DK-2200 Copenhagen N Denmark

Academic advisor: Associate Professor Sine Lo Svenningsen

Section for Biomolecular Sciences Department of Biology University of Copenhagen Denmark

Assessment committee: Anders Løbner-Olesen, Professor, (Chair)

Section for Functional genomics Department of Biology University of Copenhagen Denmark

Sanna Koskiniemi, Associate Senior Lecturer Department of Cell and Molecular Biology, Microbiology Uppsala University Sweden Jakob Møller-Jensen, Associate Professor Department of Biochemistry and Molecular Biology University of Southern Denmark Denmark

Submitted: October 27th 2014

3

Since the activator - a cell-produced chemical - seems to impose a high degree of physiological homogeneity in a pneumococcal population with respect to competence, one is forced to conclude that in this case [a] bacterial population can behave as a biological unit with considerable… coordination among its members.…One wonders whether this kind of control may not be operative in some other microbial phenomena also.

—Alexander Tomasz, 1965

4

Acknowledgements

I am tremendously grateful for having Sine Lo Svenningsen as my supervisor. She is very

inspirational and always encourages me to explore new ideas and possibilities. I always hoped, but

never imagined, that science could be this fun. Thank you. You have taught me so many invaluable

things.

I would also like to thank all the wonderful past and present members of the SLS group for good

times in- and outside the laboratory. I would especially like to thank Marianne Mortensen, Rasmus

Baadsgaard Mærkedahl and Linda Hove Christensen for their important contributions to this work.

Steen Pedersen, Michael A. Sørensen, Stanley Brown, Kim Sneppen, Anders Løbner-Olesen, Kenn

Gerdes and their groups deserve thanks for their fruitful discussions, suggestions, and a nice

working atmosphere.

I would also like to thank Thomas J. Silhavy and his group for helping me carry out a part of this

work in the Silhavy Lab, at Princeton University, and not least for how they greeted my family.

Although we snowed in three times, and Kristian busted his knee (again) it was a great experience

that we learned a lot from. I look forward to moving back with my family.

My wonderful parents deserve big thanks for always supporting me in what I want to do. From

sports, to dragging them to all kinds of events when I was too shy to go alone and to the countless

Fridays where they pick up August and give him the best time. I am grateful for the support from

the rest of my family and my friends, especially Dorte Clausen for proofreading.

Thank you Kristian, my amazing husband. I really don’t know how to express my gratitude to you,

except to say that I love you. You mean everything to me and you are the steady “anchor pier” in

my life. Thank you August, my son, for teaching me so much about love and excavators. You are

the most fun and loving boy.

This work was supported in part by the Danish National Research Foundation through the Center

for Models of Life, as well as the Novo Nordisk Foundation, and a FREJA fellowship to SLS from

the Faculty of Science, University of Copenhagen.

Nina Molin Høyland-Kroghsbo

Copenhagen, Denmark

October 2014

5

Summary Uncovering how bacteria perceive environmental signals and how they interpret these, in order to

constantly adapt to changes in their environment, is important for understanding not only microbial

ecology but also bacterial pathogenesis. Furthermore, it provides cues as to how we might interfere

with these systems, in order to prevent undesirable bacterial behavior.

In a process known as quorum sensing, bacteria emit and detect small diffusible molecules, which

upon reaching a certain extracellular concentration, activate cellular quorum sensing receptors and

thereby turn on group behavior genes. Quorum sensing controls important bacterial behaviors,

including bioluminescence, biofilm formation, and virulence. Inter- and intraspecies quorum

sensing signals enable bacteria to estimate the abundance and species complexity of a microbial

community. A long standing question in the bacterial cell-cell communication field is why E. coli

harbors SdiA, an orphan quorum sensing receptor that is activated in response to AHL quorum

sensing molecules produced by other Gram-negative species.

The overall aim of this PhD thesis was to investigate to what degree AHL quorum sensing signals

are exploited by E. coli to increase its chances of surviving potential environmental threats.

This thesis uncovers the first quorum sensing-regulated bacteriophage defense mechanism, which

serves to protect E. coli against infection by the bacteriophage viruses λ and χ. Investigating the

regulatory mechanism underlying the quorum sensing regulated defense mechanism, led to the

discovery that AHL activates expression of cnu, encoding an Hha-family protein that interacts with

the global regulatory protein H-NS, and potentially modifies its functions.

Inspired by the discovery that AHL protects E. coli from bacteriophage attacks, it was hypothesized

that AHL may be perceived by E. coli as a general signal of stress in the environment. Indeed, it

was discovered that AHL signaling upregulates the alternative sigma factor δS, the master regulator

of the general stress response in E. coli. AHL-mediated activation of the general stress response

also resulted in an increase in transiently antibiotic tolerant persister cells in E. coli.

In conclusion, this thesis provides a key answer to why E. coli listens in on AHL signals it does not

itself produce, namely that detection of interspecies AHL quorum sensing by E. coli serves to

anticipate- and adapt to environmental stresses. This discovery may have important clinical

implications, as quorum sensing-inhibitory drugs may, in addition to their primary purpose to

decrease virulence of pathogens, additionally weaken bacterial defenses thus making them prone to

succumb to a patient’s own immune system, bacteriophages, and antibiotics, and would additionally

reduce the risk of persister cell formation and thus relapse of infection after an antibiotic treatment.

6

Sammendrag (Danish summary) At afdække hvordan bakterier opfatter og fortolker signaler i deres omgivelser, for konstant at

tilpasse sig ændringer i deres miljø, er vigtigt for vores forståelse af både mikrobiel økologi og

bakteriel patogenese. Desuden vil det belyse, hvordan vi kan påvirke disse systemer for at forhindre

uønsket bakteriel adfærd. I en proces kendt som quorum sensing, udsender og detekterer bakterier

små diffundérbare molekyler, som efter at have nået en vis ekstracellulær koncentration, aktiverer

cellulære quorum sensing-receptorer og derved tænder gruppe respons gener. Quorum sensing

styrer vigtig bakteriel adfærd, herunder bioluminescens, biofilmdannelse og virulens. En

kombination af artsspecifikke og bredt benyttede quorum sensing signaler muliggører at bakterier

kan estimere tætheden og artskompleksiteten af et mikrobielt samfund. Et ubesvaret spørgsmål i det

bakterielle celle-celle kommunikations forskningsfelt er hvorfor E. coli besidder SdiA, en quorum

sensing receptor, der aktiveres af AHL quorum sensing molekyler produceret af andre Gram-

negative arter, men ikke af signaler produceret af E. coli selv.

Formålet med denne PhD afhandling var at undersøge, i hvilken grad AHL quorum sensing

udnyttes af E. coli for at øge chancerne for at overleve potentielle miljømæssige trusler. Denne

afhandling afdækker den første quorum sensing-regulerede bakteriofag forsvarsmekanisme, som

beskytter E. coli mod infektion af bakteriofag vira λ og χ. Studiet af signalvejen bag den quorum

sensing regulerede forsvarsmekanisme, førte til opdagelsen af, at AHL aktiverer ekspressionen af

cnu, der koder for et Hha-familie protein, som interagerer med det globale regulatoriske protein H-

NS, og potentielt ændrer dets funktioner. Inspireret af opdagelsen af, at AHL beskytter E. coli mod

bakteriofag angreb, fremsattes hypotesen, at AHL kan opfattes af E. coli som et generelt

stresssignal. Det blev fundet, at AHL signalering opregulerer den alternative sigmafaktor δS,

hovedregulatoren af det generelle stressrespons i E. coli. AHL-medieret aktivering af dette respons

resulterede desuden i en stigning i transient antibiotikatolerante persister celler i E. coli. Til

konklusion giver denne afhandling et essentielt svar på, hvorfor E. coli lytter til AHL-signaler den

ikke selv producerer, nemlig at AHL quorum sensing tillader E. coli at forberede sig på stress fra

omgivelserne. Denne opdagelse kan have vigtige kliniske implikationer, idet quorum sensing-

hæmmende lægemidler ud over deres primære formål at reducere virulens af patogener, desuden

kan svække bakterielle forsvarsmekanismer. Hermed ville bakterierne formegentlig blive mere

tilbøjelige til at bukke under for en patients immunsystem, for bakteriofager samt for

antibiotikabehandling. Ydermere ville reducere risikoen for dannelsen af persister celler og deraf

tilbagefald af infektion efter endt antibiotika behandling reduceres.

7

List of publication and manuscripts A quorum-sensing-induced bacteriophage defense mechanism MBio. 2013 Feb 19;4(1):e00362-12. doi: 10.1128/mBio.00362-12. Høyland-Kroghsbo NM, Maerkedahl RB, Svenningsen SL. AHL quorum sensing upregulates cnu, a member of the Hha-family of H-NS modulatory proteins in E. coli Høyland-Kroghsbo NM and Svenningsen SL. Manuscript in preparation Interspecies quorum sensing accelerates E. coli entry into stationary phase Høyland-Kroghsbo NM, Christensen LH, Svenningsen SL. Manuscript in preparation

8

List of abbreviations AHL N-acyl-L-homoserine lactone

AI Auto inducer

ChIP Chromatin immunoprecipitation

CRISPR Clustered regulatory interspaced short palindromic repeats

Cas CRISPR associated

DPD (S)-4,5-Dihydroxy-2,3-pentandione

EA Ethyl acetate

E. coli Escherichia coli

EHEC enterohemorrhagic E. coli

EMSA Electrophoretic mobility shift assay

H-NS histone-like nucleoid-structuring protein H-NS

HSL Homoserine lactone

LEE locus of enterocyte effacement

OE Over expression

ONPG o-nitrophenyl-β-D-galactopyranoside

oriC Ori gin of replication

QS Quorum sensing

SdiA Suppressor of cell division inhibitor Seq Sequencing

SD Shine-Dalgarno

S. Typhimurium Salmonella typhimurium

qRT-PCR quantitative reverse transcription PCR

9

Contents

ACKNOWLEDGEMENTS ................................................................................................... 3

SUMMARY .......................................................................................................................... 5

SAMMENDRAG (DANISH SUMMARY) .............................................................................. 6

LIST OF PUBLICATION AND MANUSCRIPTS .................................................................. 7

LIST OF ABBREVIATIONS ................................................................................................ 8

1. INTRODUCTION ........................................................................................................ 11

Aim of thesis .................................................................................................................................................................... 11

E. coli ............................................................................................................................................................................... 12

Quorum Sensing ............................................................................................................................................................. 12 A chemical language .................................................................................................................................................... 13 AHL ............................................................................................................................................................................. 13 The orphan AHL receptor SdiA ................................................................................................................................... 14 QS inhibitors ................................................................................................................................................................ 18

Bacteriophages ................................................................................................................................................................ 18 Bacteriophages and evolution of microbial communities ............................................................................................ 19

Bacteriophage λ and E. coli ......................................................................................................................................... 20

The general stress response ............................................................................................................................................ 21

The stringent response ................................................................................................................................................... 21

Persister cells ................................................................................................................................................................... 22

H-NS ................................................................................................................................................................................. 23

StpA and the Hha-family ............................................................................................................................................... 24

2. SUMMARY OF RESULTS ......................................................................................... 25

Article 1: A quorum-sensing-induced bacteriophage defense mechanism ................................................................ 25

Manuscript A: AHL quorum sensing upregulates cnu, a member of the Hha-family of H-NS modulatory proteins in E. coli ........................................................................................................................................................................... 26

Manuscript B: Interspecies quorum sensing accelerates E. coli entry into stationary phase .................................. 27

10

3. ARTICLE 1: A QUORUM-SENSING-INDUCED BACTERIOPHAGE DEFENSE MECHANISM ..................................................................................................................... 28

4. MANUSCRIPT A: AHL QUORUM SENSING UPREGULATES CNU, A MEMBER OF THE HHA-FAMILY OF H-NS MODULATORY PROTEINS IN E. COLI ............................ 42

AHL quorum sensing upregulates cnu, a member of the Hha-family of H-NS modulatory proteins in E. coli ..... 43 Abstract ........................................................................................................................................................................ 43 Introduction .................................................................................................................................................................. 43 Results .......................................................................................................................................................................... 46 AHL downregulates the mal operon proteins LamB and MalE ................................................................................... 46 AHL regulates malT posttranscriptionally ................................................................................................................... 46

AHL enhances expression of cnu ................................................................................................................................. 47 Discussion .................................................................................................................................................................... 50 Acknowledgements ...................................................................................................................................................... 51 Materials and methods ................................................................................................................................................. 51 Supplemental table S1 .................................................................................................................................................. 54 Supplemental table S2 .................................................................................................................................................. 54 References .................................................................................................................................................................... 55

5. MANUSCRIPT B: INTERSPECIES QUORUM SENSING ACCELERATES E. COLI ENTRY INTO STATIONARY PHASE................................................................................ 57

Interspecies quorum sensing accelerates E. coli entry into stationary phase ............................................................ 58

Abstract ........................................................................................................................................................................ 58 Introduction .................................................................................................................................................................. 58 Results .......................................................................................................................................................................... 60 AHL quorum sensing signaling leads to accumulation of δS ...................................................................................... 60

AHL signaling induces δS accumulation earlier in growth phase ................................................................................ 61 AHL increases persister cell formation in E. coli ........................................................................................................ 62 Discussion .................................................................................................................................................................... 63 Acknowledgements ...................................................................................................................................................... 65 Materials and Methods ................................................................................................................................................. 65 Supplemental table S1 .................................................................................................................................................. 67 References .................................................................................................................................................................... 68

6. DISCUSSION ............................................................................................................. 71

AHL as a stress-anticipation signal .............................................................................................................................. 71 Common targets for SdiA and δS ................................................................................................................................. 72 Quorum sensing-mediated regulation of phage defense .............................................................................................. 72

Cnu-mediated regulation .............................................................................................................................................. 73 Further experiments ..................................................................................................................................................... 74

7. CONCLUSION............................................................................................................ 77

8. PERSPECTIVES ........................................................................................................ 78

9. REFERENCES ........................................................................................................... 79

11

1. Introduction

The global threat of multi-drug resistant bacteria urgently demands alternatives to conventional

antibiotics. Bacteria constantly have to adapt to changes in their environment, in order to survive.

Understanding how bacteria integrate environmental signals into their strategies for survival is

critical and forms the basis upon which to develop alternative approaches to combat pathogens.

Bacterial cell-cell communication enables bacteria to assess the density and the composition of a

microbial environment. Through the orphan quorum sensing receptor SdiA, E. coli can detect AHL

quorum sensing signals emitted by other Gram-negative species in its environment. However, the

functions and extend to which E. coli makes use of the AHL-receptor SdiA to regulate its behavior

is largely unknown.

Aim of thesis

The aim of this PhD thesis was to investigate to what degree AHL quorum sensing signals are

exploited by E. coli to increase its chances of surviving potential environmental threats.

The objectives of this thesis were:

• To determine whether AHL quorum sensing regulates bacteriophage defenses in E. coli

(Paper 1).

• To unravel the regulatory pathway underlying the AHL-mediated activation of a novel

bacteriophage defense strategy in E. coli (Manuscript A).

• To investigate to what extend AHL quorum sensing signals are perceived by E. coli as a

general signal of stress in the environment (Manuscript B).

12

E. coli

Escherichia coli (E. coli) serves as a paradigm and the study of this classical model organism has

yielded many ground-breaking discoveries in biology. E. coli encompasses a Gram-negative,

facultative anaerobic, and rod-shaped group of bacteria belonging to the genus Escherichia. E. coli

is an extraordinarily diverse species with a high degree of both genetic and phenotypic diversity

(reviewed in (Chaudhuri and Henderson, 2012; Leimbach et al., 2013)). E. coli is an important

commensal bacterium of the mammalian intestinal microflora. However, certain pathogenic strains

cause serious human illness, namely enteric- and diarrheal disease, urinary tract infections, sepsis,

and meningitis (reviewed in (Kaper et al., 2004)).

Quorum Sensing

Bacteria have developed sophisticated mechanisms that allow them to perceive and respond to ever-

changing environmental conditions, which is crucial for their survival and ability to colonize

extremely diverse ecological niches (reviewed in (Pereira et al., 2013)).

Quorum sensing (QS) is the process of bacterial cell-cell communication via emission and detection

of small diffusible molecules called autionducers (AIs). Bacteria constantly produce AI’s and

accordingly, at low-cell-density AI levels are low, whereas AI’s accumulate at high-cell-density.

This enables bacteria to assess cell density and adjust gene expression accordingly, thus allowing

bacteria to behave as a multicellular organism. Upon reaching an AI threshold level, AI’s activate

bacterial QS receptors and consequently group-behavior genes are simultaneously turned on

throughout the population. Thereby group-behavior processes are activated only when adequate

numbers of cells unite to efficiently perform the task (reviewed in (Ng and Bassler, 2009)). The

phenomenon of QS was first observed when a self-produced, extracellular factor was found to

regulate Streptococcus pneumonia entry into the competent state (Tomasz, 1965). Later, the

phenomenon was described in the regulation of bioluminescence in the marine bacterium Vibrio

fischeri (Nealson et al., 1970). Vibrio fischeri forms a symbiotic relationship with the squid

Euprymna scolopes where it colonizes the nutrient rich light organ, grows to high cell densities and

turns on expression of its luciferase operon in response to accumulating AI’s. QS also regulates

13

important bacterial group-behaviors including virulence and biofilm formation (reviewed in (Ng

and Bassler, 2009)).

A chemical language Bacteria communicate via different chemical languages dependent on species. QS signaling occurs

within and between species and even interkingdom signaling has been discovered (reviewed in

(Walters and Sperandio, 2006)). As part of their warfare, bacteria can spread misinformation by

sequestering or even degrading signals by expressing lactonases, thus muting their competitors

(reviewed in (Bassler and Losick, 2006; Smith et al., 2011)).

There are two typical classes of QS AIs in Gram-negative bacteria: N-acyl-L-homoserine lactones

(AHLs) and a family of spontaneously interconverting molecules formed from (S)-4,5-Dihydroxy-

2,3-pentandione (DPD), collectively known as AI-2. AHLs are used for intraspecies

communication while AI-2 is a universal signal for interspecies communication (reviewed in (Ng

and Bassler, 2009; Walters and Sperandio, 2006)). Production of AI-2 is widespread among the

bacterial kingdom and is found in diverse species of both Gram-negative and Gram-positives

(reviewed in (Pereira et al., 2013)).

E. coli produces indole and it has been proposed to serve as an AI for activation of SdiA and

furthermore SdiA has been found to be required for indole-mediated inhibition of biofilm formation

(reviewed in (Lee and Lee, 2010)), however it was later found that SdiA does not positively respond

to indole but is rather inhibited by it (Sabag-Daigle et al., 2012).

AHL AHLs are a major class of intraspecies AIs produced

by Gram-negative species. AHLs are composed of a

homoserine lactone ring with an acyl side chain

(Figure 1). The length of the acyl chain varies from C4

to C18, and can additionally be modified at C3. The

nature of the acyl side chain specifies the class of

AHL, where typically only one type of AHL is

produced and detected by a single species (reviewed

Figure 1. Structure of a homoserine lactone ring, where variations in acyl chains (R) specifies the class of AHL. Acyl chains range in length from C4 to C18, where C3 can be further modified. Adapted from (Ng and Bassler, 2009).

14

in (Ng and Bassler, 2009)). The freely diffusible AHLs are synthesized by members of the LuxI

family of synthases and are detected by LuxR type receptors, which are transcription factors that

regulate group-behavior genes. luxI/luxR homologues are widely distributed among Gram-negative

species (reviewed in (Soares and Ahmer, 2011)).

The orphan AHL receptor SdiA Species in the genera Escherichia, Salmonella, Klebsiella and Shigella harbor the orphan luxR

homologue SdiA but not a cognate synthase. Thus, these species are able to listen in on AHL QS

emitted by other Gram-negative species (reviewed in (Smith et al., 2011)). E. coli and Salmonella

typhimurium (S. Typhimurium) SdiA responds to a wide range of AHLs emitted by other Gram-

negative species, but it is especially responsive to the AHL variants oxo-C6-HSL and oxo-C8-HSL

(Michael et al., 2001; Soares and Ahmer, 2011). SdiA contains a C-terminal helix-turn-helix DNA

binding domain and an N-terminal AHL-binding domain (Wang et al., 1991; Yao et al., 2006). In

the absence of AHL’s, E. coli SdiA expressed from a plasmid, forms insoluble inclusion bodies.

AHL binding to SdiA induces a folding switch which makes SdiA soluble (Yao et al., 2006).

SdiA expression appears to be QS-regulated through the AI-2 QS system. The effects of deleting

luxS, encoding the AI-2 synthase, was investigated in a microarray study of enterohemorrhagic E.

coli (EHEC) and in the luxS mutant, sdiA levels were downregulated 11 fold as compared to the WT

strain (Sperandio et al., 2001). Interestingly, SdiA is also involved in regulation of the AI-2 QS

system, which suggests that there is a regulatory loop between the two QS systems in E. coli (Zhou

et al., 2008). The crosstalk between these two QS systems may allow E. coli to fine tune its QS

perception in response to the population distribution of AHL-producing Gram-negative species and

the population density of both closely- and more distantly related species, through AI-2 signaling.

In S. Typhimurium, sdiA was found to be expressed in a bimodal fashion during the growth phase,

with a burst in exponential phase and again in stationary phase, where the stationary phase sigma

factor δS was responsible for the latter. The authors further found the cAMP receptor protein Crp to

be the major direct activator of sdiA expression in exponential and early stationary phase. LeuO was

found to be a minor activator of sdiA and the LeuO binding site in the sdiA promoter was highly

conserved among Enterobacteriaceae (Turnbull et al., 2012).

15

In an electrophoretic mobility shift assay (EMSA) based screen of 191 of approximately 300 E. coli

transcription factors, 15 were found to bind the sdiA promoter, including SdiA itself. Five of these

were further tested and the osmoregulated two component system regulator OmpR showed the

highest regulatory potential on sdiA. ompR deletion upregulated sdiA promoter fusions to lacZ and

lux reporter genes compared to WT and overexpression of ompR complemented the ompR mutant

phenotype. Northern blot analysis showed that ompR deletion led to a high accumulation of sdiA

mRNA, suggesting that OmpR is a transcriptional repressor of sdiA (Shimada et al., 2013).

Interestingly, in S. Typhimurium, sdiA expression was previously found to be osmoregulated,

pointing towards common regulators of sdiA expression between E. coli and S. Typhimurium

(Turnbull et al., 2012).

Controversy exists regarding SdiA-mediated regulation as there are major inconsistencies in

reported SdiA targets, some of which may be explained by differences in the methods used to

identify targets and may be further complicated by the growth phase-dependent expression of SdiA

(Turnbull et al., 2012). Generally these studies are either performed using plasmid-based over

expression of sdiA or chromosomally encoded sdiA in the presence or absence of the activating

AHL QS signals. Table 1 Lists selected SdiA targets and the experimental methods used to identify

them.

16

Method Genes or processes regulated by SdiA Reference

Up regulated Down regulated

Unaffected

Response to sdiA over expression

(OE) and deletion, 37 °C

ftzQAZ (Wang et al., 1991)

Response of lacZ reporter fusions

to C10-HSL, oxoC6-HSL, OHC4-HSL in sdiA deletion or sdiA OE in the sdiA deletion strain, 28 °C

ftsQA (Sitnikov et al., 1996)

Response to sdiA OE in EHEC, 37

°C

ftsZ fliC, espD, eae (Kanamaru et al., 2000)

Microarray of sdiA OE vs. WT, 37

°C

sdiA, ftzQAZ,

acrABDEF

Flagellar genes,

malE

(Wei et al., 2001)

Promotor trap response to C6-

HSL in WT, 30 °C and 37 °C

gadA

(30 °C)

malT

(30 °C)

gadA, malT

(37 °C)

(Van Houdt et al., 2006)

Response to oxoC6-HSL in WT,

37 °C

ydiV (Zhou et al., 2008)

Response of sdiA deletion vs. WT

EHEC, 37 °C

fliC, csgA LEE genes (Sharma et al., 2010)

Screen of luciferase fusions and

selected targets in WT

responding to oxoC6-HSL, 30 °C

and 37 °C

gadW, gadE

(30 and 37 °C)

fliE

(30 and 37 °C)

ftzQAZ, acrAB

(30 and 37 °C)

(Dyszel et al., 2010)

qRT-PCR of EHEC in response to

oxoC6-HSL, 37 °C

gadX, gadW LEE genes (Hughes et al., 2010)

Table 1. An overview of selected reported targets of SdiA-mediated regulation in E. coli. Important experimental conditions are noted in the first column. Unless otherwise stated, the studies were carried out using E. coli K12. Genes and processes that are found to be either up- or downregulated, or unaffected, by SdiA are noted. The list of genes found regulated is not exhaustive, as mainly genes of special interest to this thesis are noted and which have been functionally tested or, as in the case of the microarray study, entire functional groups that are regulated in a similar fashion.

The SdiA gene was first isolated on the basis of its ability to bypass inhibition of cell division, when

over expressed and was therefore termed, “Suppressor of cell division inhibitor” (Wang et al.,

1991). Overexpression of SdiA was found to increase transcription from the P2 promoter of the

essential cell-division gene cluster ftsQAZ, which led to increased FtsZ protein levels. Deletion of

sdiA inhibited P2-dependent transcription, but had no apparent effect on cell growth or division,

suggesting that sdiA is dispensable for growth under the standard laboratory conditions applied

(Wang et al., 1991). The later reports were also based on overexpression of SdiA rather than

chromosomally encoded sdiA, and they supported the findings by Wang et al. (Kanamaru et al.,

2000; Sitnikov et al., 1996; Wei et al., 2001).

17

Importantly, Dyszel et al. determined that the previously reported targets regulated by plasmid-

expressed SdiA, ftsQAZ and acrAB do not respond to AHL signaling through SdiA when it is

expressed from its chromosomal location, they tested this at both 30 °C and 37 °C. Using a

luciferase based fusion library, they identified genes from the glutamate-dependent acid resistance

system upregulated and fliE downregulated by chromosomally expressed SdiA (Dyszel et al.,

2010).

The study by Dyszel et al. questions the biological relevance of SdiA-mediated regulation of the

defining targets of SdiA, i.e. ftsQAZ. Although SdiA has been found to bind the ftzQ promoter

directly (Hughes et al., 2010; Kanamaru et al., 2000; Yamamoto et al., 2001), it may not follow that

SdiA significantly regulates ftzQAZ transcription, at least not under the experimental conditions

used by Dyszel et al. One could imagine that certain environmental signals could upregulate SdiA

to an extent where it would turn on ftzQAZ. Alternatively, the discrepancies regarding SdiA-

mediated regulation may also be explained by the growth phase-dependent regulation of SdiA

expression (Turnbull et al., 2012), which inherently will lead to a growth phase- dependent capacity

of SdiA to mediate regulation, as observed by us and others (Dyszel et al., 2010). The study by

Dyszel et al. highlights the importance of experimental considerations, when investigating SdiA

functions. Of notice, SdiA expressed from its chromosomal location shows a regulatory potential on

a subset of targets, even in the absence of activating AHL signals (Hughes et al., 2010; Zhou et al.,

2008), suggesting that it has some inherent activity on specific targets.

A SELEX-chip screen, based on several rounds of immunoprecipitation of SdiA in the presence of

E. coli genomic DNA fragments, followed by microarray analysis, sought to identify direct targets

of SdiA (Shimada et al., 2014). The authors identified targets that were bound by SdiA in the

absence- or in the presence of C4-HSL or synthetic HSL analogues, none of which were the AHL

molecules most potent to activate SdiA (Michael et al., 2001; Soares and Ahmer, 2011). Northern

blot validation of targets identified by the SELEX-chip method did not overlap with any of the

previously identified SdiA targets (Shimada et al., 2014).

If considering only the studies of SdiA expressed from its chromosomal location, a general picture

emerges where SdiA in E. coli K-12 and EHEC inhibits flagellar synthesis (Dyszel et al., 2010;

Sharma et al., 2010) and virulence genes in the locus of enterocyte effacement (LEE) pathogenicity

island of enterohemorrhagic E. coli (EHEC) (Hughes et al., 2010) and positively regulates

18

glutamate-dependent acid resistance genes (Dyszel et al., 2010; Hughes et al., 2010; Van Houdt et

al., 2006).

QS inhibitors Inhibitors of bacterial cell-cell communication are emerging as promising alternatives to

conventional antibiotics in the battle against multi-drug resistant bacteria. The reasoning is that as

QS is required for virulence of many pathogens, disrupting bacterial cell-cell communication may

render these harmless, allowing patient’s immune systems to clear infections. This approach,

aiming at disarming rather than killing pathogens, would additionally minimally disturb the natural

microbiome. Moreover, it would put less selective pressure on the bacteria, which would reduce the

risk of resistant populations emerging (reviewed in (Marx, 2014; Rutherford and Bassler, 2012)).

Approaches to interfere with QS include antagonizing QS receptors, inhibition of AI production and

degrading- and trapping AI’s (reviewed in (Hirakawa and Tomita, 2013)).

Bacteriophages Bacteriophages (phages) are the most abundant biological entity on Earth. They are viruses that

infect and replicate inside bacteria or archea and they are estimated to outnumber their bacterial

hosts tenfold in most natural environments. Phage genomes are exceptionally diverse and they carry

an enormous pool of genes with unknown function and often no homology to previously sequenced

genes (reviewed in (Brussow and Hendrix, 2002)). Phages λ and χ belong to the order

Caudovirales, which encompasses the tailed phages with double-stranded DNA packed into a

polyhedral head. It is estimated that the number of tailed phages in the biosphere amounts to 1031

(Veesler and Cambillau, 2011).

Phages can be either virulent or temperate. Upon infecting its bacterial host, a virulent phage begins

to replicate and finally lyses its host to release phage progeny. A temperate phage can alternate

between a lytic state, where it immediately starts producing offspring and lyse its host upon

infection, or it can form a lysogen by integrating phage DNA into the host chromosome, where it is

replicated along with its host as a prophage. Upon certain changes in the metabolic state and stress

level of the lysogenic host cell, some prophages can switch to enter the lytic state (Little and Mount,

1982; Melechen and Go, 1980; Ptashne, 1986). Cryptic prophages are defective prophages that have

19

lost the ability to enter the lytic state. They often confer benefits to their host by providing

immunity against specific infecting phages or by expressing virulence factors. Moreover, phages

play an important role in horizontal gene transfer, enabling transfer of genetic material between host

bacteria, including toxins and antibiotic resistance genes. Additionally, bacteria have adopted entire

phage tail modules and use these to produce phage-tail-like bacteriocins, which are antimicrobial

compounds that are effective against closely related bacterial species (reviewed in (Veesler and

Cambillau, 2011)).

Bacteriophages and evolution of microbial communities

There is an ongoing evolutionary arms race between bacteria and their predators, the bacteriophage

viruses (reviewed in (Samson et al., 2013)). The predation pressure by phages is substantial and

consequently bacteria have evolved a broad spectrum of diverse anti-phage strategies, ranging from

blocking initial attachment over intricate ways to degrade the incoming phage genome to abortive

infection, where host suicide prevents spread of phage progeny into the population (reviewed in

(Labrie et al., 2010)). A fascinating example of an antiphage system, which has additionally

revolutionized eukaryotic genome editing, is the Clustered regulatory interspaced short palindromic

repeats (CRISPR) system with its CRISPR associated (Cas) genes. The CRISPR-Cas adaptive

microbial immune system provides bacteria and archaea with acquired and heritable sequence-

specific immunity against previously encountered phages and plasmids (reviewed in (Pennisi, 2013;

Westra et al., 2012)).

Phages, in turn, have evolved mechanisms to bypass all these defenses. Examples include phages

that can enzymatically degrade the protective exopolysaccharide shield produced by some bacteria,

and thereby gain access to the buried phage receptors. Other phages can shuffle their tail specificity

and thereby alter their receptor- and host range. Phages can even encode their own CRISPR-Cas

systems and target these against the bacterial host (reviewed in (Samson et al., 2013)). Moreover,

some phages appear to have the ability to interfere with bacterial QS. A phage that encodes an AI

synthase has been found, and a database search identified multiple phage genomes harboring

receptors for these molecules, indicating that phages may disturb bacterial cell-cell signaling and

even listen in on the signals (Hargreaves et al., 2014). A phage capable of listening in on bacterial

communication would be able to quantify the abundance of potential prey, as high levels of

20

bacterial cell-cell signals accumulate at high-cell-density (reviewed in (Ng and Bassler, 2009)). This

could in turn be exploited by the phage to launch an attack when the potential prey is most abundant

and hence the probability to spread was highest. The observation that incubation of soil or ground

water samples with synthetic AHLs resulted in up to a 14-fold increase in the occurrence of free

phage particles may be the result of QS-mediated induction of bacteriophage replication or it may

be due to QS-activated phage defenses, resulting in accumulations of free phages, or a combination

of both (Ghosh et al., 2009).

The interactions between bacteria and their phages shape microbial communities. Since maintaining

constantly activated defenses places a metabolic burden on bacteria, evolution likely selects against

this (Gomez and Buckling, 2011; Hall et al., 2011). If bacteria could instead regulate their defenses

dynamically in response to estimates of current risk of phage attacks, they would have a competitive

advantage.

Bacteriophage λ and E. coli The coliphage λ was discovered more than 60 years ago (Lederberg, 1950), and it has since become

a paradigm for studying key biological mechanisms including gene regulation, developmental

pathways, recombination, molecular cloning, protein folding, and virion assembly (reviewed in

(Court et al., 2007; Gottesman and Weisberg, 2004)). Consequently, λ phage is considered the most

completely understood biological entity, yet new discoveries of λ biology and its interactions with

its host E. coli still emerge. One of the reasons that λ has gained so much interest is that it is a

temperate phage and therefore has two developmental paths that it can choose from (Lieb, 1953a,

b). As such the λ genetic switch has become a paradigm for decision making gene circuits in simple

organisms (reviewed in (Oppenheim et al., 2005)).

LamB is the maltoporin used to transport long chain maltodextrins across the outer membrane

(reviewed in (Boos and Shuman, 1998)) and it is utilized by λ as a receptor for attachment (Randall-

Hazelbauer and Schwartz, 1973). lamB is encoded in the malB region of the maltose regulon, which

is under positive regulation by MalT (reviewed in (Boos and Shuman, 1998)).

21

The general stress response The rpoS gene encodes the stationary phase alternative sigma factor δS, which is the master

regulator of the general stress response. δS associates with RNA polymerase and promotes

transcription of the rpoS regulon, which includes up to 10% of the E. coli genome (reviewed in

(Landini et al., 2014)). This major shift in gene expression, favoring genes required to cope with a

wide range of stresses, is the hallmark of the general stress response. The general stress response

renders cells resistant or tolerant to multiple stresses, in contrast to the specific stress responses,

which are triggered by a single stress signal that in turn allows cells to cope with this specific stress

situation only (reviewed in (Hengge-Aronis, 2002)).

δS accumulates upon entry into stationary phase and in response to other stress signals, including

changes in temperature, pH, osmolarity and nutrient availability, which is often accompanied by a

reduction in growth rate. rpoS is regulated at the level of transcription initiation, mRNA stability

and translation, and the activity and stability of the δS protein is also regulated (Brescia et al., 2004)

and (reviewed in (Battesti et al., 2011)).

The stringent response

The stress alarmone (p)ppGpp is the central regulatory molecule in the stringent response, which is

a stress response to amino acid starvation. Amino acid starvation results in accumulation of

uncharged tRNAs in the cell, which in turn activates (p)ppGpp synthesis by RelA. SpoT is

responsible for ppGpp synthesis in response to other stress signals. (p)ppGpp accumulation leads to

a redirection of transcription, favoring genes important for survival during starvation and virulence

genes, and inhibits components of the transcriptional and translational apparatus thus inhibiting

growth. Specifically, (p)ppGpp is involved in regulating both the production and activity of δS

(reviewed in (Magnusson et al., 2005; Starosta et al., 2014)).

Due to the lack of (p)ppGpp, a relA spoT strain expresses very little δS and consequently has

multiple amino acid requirements and is hyper sensitive to stress including heat and cold shock,

osmotic stress, and antibiotics (Hengge-Aronis, 2002; Maisonneuve et al., 2013; Xiao et al., 1991).

22

Persister cells

Both the stringent response and the general stress response are involved in formation of persister

cells (Korch et al., 2003; Maisonneuve et al., 2013; Tkachenko et al., 2014), which are a transiently

multidrug tolerant, slow growing subpopulation of an isogenic bacterial population (reviewed in

(Helaine and Kugelberg, 2014; Maisonneuve and Gerdes, 2014)).

It has been proposed that the stringent response may form the basis of the persistent stage (Korch et

al., 2003). The model for (p)ppGpp-mediated formation of persisters is based on stochastic

induction of (p)ppGpp (Maisonneuve et al., 2013), which inhibits exopolyphosphatase, the cellular

enzyme that degrades polyphosphate. The resulting increase in polyphosphate activates the Lon

protease to degrade all type II antitoxins. The freed toxins then inhibit translation and cell growth

and thereby result in persisters (reviewed in (Maisonneuve and Gerdes, 2014))

Maisonneuve et al. used an rpoS-mCherry reporter as an indicator for (p)ppGpp levels to study

single-cell persister cells and they found that persisters are stochastically formed. Upon transfer of

an exponentially growing culture to a microfluidic device, persister cells were identified based on

mCherry expression and slow growth and they were monitored by time lapse microscopy during

subsequent antibiotic treatment, showing that the non-persisters were killed and upon addition of

antibiotic-free media, the persisters were able to re-grow (Maisonneuve et al., 2013). Maisonneuve

et al. also demonstrated that E. coli in stationary phase have an approximately 100-fold higher

persister fraction compared to exponentially growing E. coli (Maisonneuve et al., 2011).

Tkachenko et al. recently demonstrated that δS induced persister cell formation in E. coli. They

found that ectopic expression of rpoS in exponentially growing E. coli led to a more than tenfold

increase in cells persistent to the antibiotic netilmicin and they further showed that stationary phase

WT cells had a 100-fold higher persister frequency compared to an isogenic rpoS mutant

(Tkachenko et al., 2014).

Contrary to this, Hong et al. found that deletion of rpoS led to almost complete resistance to

ampicillin (Hong et al., 2012). They argued that the persisters they observe are not formed prior to

addition of antibiotics, but are rather formed in response to the antibiotics.

23

Persister cells are believed to cause recurrence of infections after antibiotic treatment. Since relapse

of infections pose a major medical challenge to the health care systems, many efforts to eliminate

persisters are currently undertaken (reviewed in (Helaine and Kugelberg, 2014)).

H-NS The histone-like nucleoid-structuring protein (H-NS) acts as a global transcriptional silencer in

gram-negative bacteria. The H-NS protein is highly abundant and is present in about 20,000

monomers pr. cell (reviewed in (Dorman, 2014)). It preferentially binds and silences intrinsically

curved AT-rich DNA, yet a few cases of H-NS-mediated regulation of RNA targets have been

described (Brescia et al., 2004; Park et al., 2010).

The most well described function of H-NS is its role in silencing horizontally acquired genes,

including virulence genes in E. coli and S. Typhimurium. It is believed that modulators (described in

the next section) of H-NS activity allows expression of these genes under specific circumstances

(reviewed in (Dorman, 2014; Stoebel et al., 2008)).

H-NS has the highest affinity for curved dsDNA but also associates with specific RNAs with a 3-4

fold lower binding affinity compared to curved DNA (Atlung and Ingmer, 1997; Brescia et al.,

2004). Although H-NS is a global transcriptional silencer, it acts as a translational enhancer on a set

of mRNAs containing weak Shine-Dalgarno sequences (Park et al., 2010). H-NS was found to

enhance translation of malT, lrhA, dpiA and znuA mRNAs. The mechanism was unraveled for malT,

where H-NS was found to bind directly to the AU rich stretch -35 to -40 relative to the start codon

of malT and thereby repositioning the 30S pre-initiation complex in a more favorable position for

translation (Park et al., 2010).

H-NS is composed of an N-terminal oligomerisation domain attached via a flexible linker to a C-

terminal nucleic acid binding domain. This structure enables H-NS dimers to form wide-ranging

DNA-protein-DNA bridges that impede the movement of RNA polymerase, and thus silences the

bound DNA regions (reviewed in (Stoebel et al., 2008)).

24

StpA and the Hha-family

The StpA protein is a close paralogue of H-NS and it is structurally highly similar to H-NS. StpA is

approximately sevenfold less abundant than H-NS (Wang et al., 2012), it is able to form

heterodimers with H-NS, and it can partially compensate for the loss of H-NS (reviewed in (Stoebel

et al., 2008)).

Cnu (also known as YdgT) and its close paralogue Hha belong to a family of small proteins that are

homologous to the oligomerization domain of H-NS but lack the DNA-binding domain. Cnu and

Hha are the only two chromosomally encoded members of the Hha family in E. coli (reviewed in

(Madrid et al., 2007)). Cnu and Hha are approximately 130 and 5.000 fold less abundant than H-NS

respectively (Wang et al., 2012).

The extensive interactions of H-NS, StpA, Hha and Cnu were investigated by Paytubi et al. Cnu

and Hha were found to form complexes with both H-NS and StpA in vitro and Hha formed a

complex with StpA. No direct interaction was found between Cnu and Hha. The interaction of Hha

and Cnu with StpA protected StpA from proteolysis. In a hha mutant, hns expression was decreased

whereas cnu expression was increased, and the latter partly compensated for the loss of hha. In a

hha cnu double mutant, both hns and stpA were downregulated (Paytubi et al., 2004).

Hha forms heteromers with H-NS and thereby alters the target specificity of H-NS (reviewed in

(Dorman, 2014)). It has been proposed that Hha-family proteins generally have evolved to fine-tune

the regulatory activity of H-NS and StpA on specific sets of environmentally modulated genes

(Madrid et al., 2007; Paytubi et al., 2004)

25

2. Summary of results This PhD thesis is based on one published article and two manuscripts, for which further

experiments are needed before they will be submitted for publication.

Article 1: A quorum-sensing-induced bacteriophage defense mechanism This study investigates how QS affects phage-host interactions in the classical model system of

phage λ and E. coli.

Since phages require a bacterial host for proliferation, they can rapidly spread throughout a high-

cell-density population of susceptible bacteria. Therefore, in high-cell-density situations, bacterial

populations are particularly vulnerable to viral infections. As noted in the introduction, bacteria

have developed anti-viral defense mechanisms. However, maintaining constantly elevated defenses,

irrespective of risk, is costly in terms of energy and likely evolution selects against this behavior

(Hall et al., 2011). We hypothesized that bacteria may use QS in their risk assessment of phage

infection and regulate their defenses accordingly.

In this article we describe the first QS-regulated phage defense mechanism. Specifically, we find

that, in response to AHL AIs, E. coli downregulates LamB, the outer membrane protein exploited

by λ to infect E. coli, thus rendering a fraction of the bacteria transiently tolerant to λ. This is

evident by reduced rates of adsorption of phage particles to the bacterium, which in turn leads to a

four-fold increase in E. coli cells surviving a potent attack by virulent λ. We additionally find that

AHL reduces adsorption of the broad-host-range phage χ to E. coli, and we propose that QS control

of phage susceptibility may be a general phenomenon in microbial communities.

26

Manuscript A: AHL quorum sensing upregulates cnu, a member of the Hha-family of H-NS modulatory proteins in E. coli We sought to investigate the regulatory mechanism responsible for the AHL-mediated down-

regulation of LamB and the consequent increase in λ tolerance.

In this study we find that AHL downregulates both LamB and MalE, belonging to the MalT

regulon. We show that malT is downregulated posttranscriptionally in response to AHL. H-NS has

previously been found to enhance translation of malT (Park et al., 2010). We do not find an effect of

AHL on H-NS expression, but we find that AHL enhances expression of cnu, an Hha-family protein

that is known to interact with H-NS (Paytubi et al., 2004). Except form the finding that cnu

expression was increased in an hha mutant background (Paytubi et al., 2004), SdiA is to the best of

our knowledge the first described regulator of cnu expression.

We argue that the AHL-mediated upregulation of cnu may modulate the H-NS-mediated

translational enhancement of malT, the positive regulator of the mal operon, including lamB.

Although overexpression of Cnu in the absence of AHL mimics the negative AHL-effect on malT,

the AHL-mediated regulation of cnu cannot fully explain the AHL-effect on a translational malT-

lacZ fusion, as AHL is still able to affect this reporter to some degree in a cnu and a cnu hha mutant

background. Thus, it appears that AHL-mediated regulation of malT can occur via two or more

redundant pathways.

27

Manuscript B: Interspecies quorum sensing accelerates E. coli entry into stationary phase

Inspired by our discovery that E. coli utilizes AHL QS signals to anticipate environmental stress in

the form of attack by phages, we hypothesized that E. coli may use QS signals to activate more

general defenses against environmental stresses. The appearance of high densities of other bacteria

could signify other stresses such as competition for space and/or nutrients, exposure to microbially

produced toxins, or entry into a different environment.

In this manuscript we report that AHL signaling leads to accelerated accumulation of δ

S, the sigma

factor induced upon entry into stationary phase. δS is responsible for activating expression of a

broad range of stress-resistance genes required in order to resist stationary phase stress (reviewed in

(Landini et al., 2014)). Specifically, we find that AHL modulates the timing of δS induction by

inducing accumulation of δS earlier in the growth phase. We also investigate the effect of AHL on

E. coli persistence to ampicillin as an δS-regulated environmental stress response. We find that AHL

increases persister cell formation in E. coli and this may be a consequence of the AHL-induction of

δS, as stationary phase cells are known to contain a higher fraction of persister cells (Maisonneuve

et al., 2013).

28

3. Article 1: A quorum-sensing-induced bacteriophage defense mechanism

doi:10.1128/mBio.00362-12. 4(1): .mBio. Mechanism

A Quorum-Sensing-Induced Bacteriophage Defense2013. Mærkedahl and Sine Lo SvenningsenNina Molin Høyland-Kroghsbo, Rasmus Baadsgaard  Bacteriophage Defense MechanismA Quorum-Sensing-Induced

http://mbio.asm.org/content/4/1/e00362-12.full.htmlUpdated information and services can be found at:

MATERIALSUPPLEMENTAL http://mbio.asm.org/content/4/1/e00362-12.full.html#SUPPLEMENTAL

REFERENCES

http://mbio.asm.org/content/4/1/e00362-12.full.html#ref-list-1This article cites 48 articles, 24 of which can be accessed free at:

CONTENT ALERTS

more>>article), Receive: RSS Feeds, eTOCs, free email alerts (when new articles cite this

  http://journals.asm.org/subscriptions/To subscribe to another ASM Journal go to:

http://mbio.asm.org/misc/contentdelivery.xhtmlInformation about Print on Demand and other content delivery options:

http://mbio.asm.org/misc/reprints.xhtmlInformation about commercial reprint orders:

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

A Quorum-Sensing-Induced Bacteriophage Defense Mechanism

Nina Molin Høyland-Kroghsbo, Rasmus Baadsgaard Mærkedahl, Sine Lo Svenningsen

Institute of Biology, University of Copenhagen, Copenhagen, Denmark

ABSTRACT One of the key determinants of the size, composition, structure, and development of a microbial community is thepredation pressure by bacteriophages. Accordingly, bacteria have evolved a battery of antiphage defense strategies. Since main-taining constantly elevated defenses is costly, we hypothesize that some bacteria have additionally evolved the abilities to esti-mate the risk of phage infection and to adjust their strategies accordingly. One risk parameter is the density of the bacterial pop-ulation. Hence, quorum sensing, i.e., the ability to regulate gene expression according to population density, may be animportant determinant of phage-host interactions. This hypothesis was investigated in the model system of Escherichia coli andphage �. We found that, indeed, quorum sensing constitutes a significant, but so far overlooked, determinant of host susceptibil-ity to phage attack. Specifically, E. coli reduces the numbers of � receptors on the cell surface in response to N-acyl-L-homoserinelactone (AHL) quorum-sensing signals, causing a 2-fold reduction in the phage adsorption rate. The modest reduction in phageadsorption rate leads to a dramatic increase in the frequency of uninfected survivor cells after a potent attack by virulent phages.Notably, this mechanism may apply to a broader range of phages, as AHLs also reduce the risk of � phage infection through adifferent receptor.

IMPORTANCE To enable the successful manipulation of bacterial populations, a comprehensive understanding of the factors thatnaturally shape microbial communities is required. One of the key factors in this context is the interactions between bacteria andthe most abundant biological entities on Earth, namely, the bacteriophages that prey on bacteria. This proof-of-principle studyshows that quorum sensing plays an important role in determining the susceptibility of E. coli to infection by bacteriophages �and �. On the basis of our findings in the classical Escherichia coli-� model system, we suggest that quorum sensing may serveas a general strategy to protect bacteria specifically under conditions of high risk of infection.

Received 17 September 2012 Accepted 16 January 2013 Published 19 February 2013

Citation Høyland-Kroghsbo NM, Mærkedahl RB, Svenningsen SL. 2013. A quorum-sensing-induced bacteriophage defense mechanism. mBio 4(1):e00362-12. doi:10.1128/mBio.00362-12.

Editor Sankar Adhya, National Cancer Institute

Copyright © 2013 Høyland-Kroghsbo et al. This is an open-access article distributed under the terms of the Creative Commons Attribution-Noncommercial-ShareAlike 3.0Unported license, which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

Address correspondence to Sine Lo Svenningsen, SLS@bio.ku.dk.

Many bacterial species possess the ability to regulate gene ex-pression according to the cell density of the population.

They do so by releasing signaling molecules called autoinducersinto the environment. As the population density increases, auto-inducers accumulate extracellularly, and this can be detected bythe bacteria—a phenomenon known as quorum sensing. Sincequorum-sensing signals vary between different bacterial species,they can be utilized to distinguish groups of bacteria. Hence, quo-rum sensing enables sensing of local population density as well asspecies complexity and makes it possible for microorganisms toswitch between different gene expression patterns depending onthese parameters. Accordingly, quorum sensing is often used tocoordinate social behaviors such as virulence and biofilm forma-tion across the population (reviewed in reference 1). This studywas motivated by the hypothesis that quorum sensing could ad-ditionally be used as a means of regulating phage-bacterium in-teractions.

Bacteriophages are viruses that attack bacteria. The predationpressure by phages is substantial, as they outnumber bacterial cellsby an estimated 10-fold in many natural environments (2). Con-sequently, host bacteria have evolved a wide range of antiphagemechanisms, including ways of blocking the initial attachment of

phages, degradation of the phage genome, or abortive infection byhost suicide, preventing the spread of phage progeny in the pop-ulation (reviewed in reference 3). As phages require a bacterialhost to proliferate, phages are expected to be more abundant anddiverse in densely populated mixed-species environments than insparsely populated environments. Therefore, the risk of sufferingphage attacks is generally elevated at high microbial cell densities.The costs associated with general phage resistance mechanismsare substantial and serve as a key factor in shaping the evolution-ary dynamics between the phage and host (4, 5). We speculate thatif bacteria used quorum sensing to regulate their antiphage activ-ities, they could specifically upregulate their defense mechanismsto avoid infection during growth under high-risk conditions,while saving the metabolic burden of maintaining constantly ele-vated antiphage strategies.

As a first approach to testing whether quorum sensing is usedto regulate phage-bacterium interactions, we investigated the roleof quorum sensing in the classical model system of phage � and itshost Escherichia coli K-12. Since the discovery of phage � morethan 60 years ago (6), it has been intensively studied and likelyrepresents the most completely understood biological entity. Theinvestigations of � and its interactions with E. coli have served as a

RESEARCH ARTICLE

January/February 2013 Volume 4 Issue 1 e00362-12 ® mbio.asm.org 1

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

paradigm for molecular biology and paved the way for our under-standing of key biological mechanisms, including gene regulation,recombination, molecular cloning, protein folding, and virion as-sembly (reviewed in reference 7).

Gram-negative bacteria typically quorum sense through theproduction and detection of N-acyl-L-homoserine lactone (AHL)quorum-sensing signals. AHLs are produced by synthases of theLuxI family and typically detected by receptors of the LuxR family(reviewed in reference 1). E. coli can detect AHL quorum-sensingsignals through SdiA, a LuxR-type transcriptional regulator (8, 9).SdiA is able to bind and be activated by a broad range of AHLs (10,11). Curiously, E. coli and other enterobacteria are not able toproduce AHLs, as they lack a LuxI-type AHL synthase, but theycan detect AHLs emitted by other bacterial species (10). Only a fewgene groups have reproducibly been shown to be regulated byAHLs in E. coli. These include flagellar genes, acid resistance genes,and virulence genes in the LEE pathogenicity island of enterohe-morrhagic E. coli (EHEC) (12–14).

In this study, we have identified a novel quorum-sensing-regulated antiphage defense mechanism in E. coli K-12. We findthat E. coli utilizes AHL quorum-sensing signals to reduce its sus-ceptibility to infection by phage � as well as the broad-host-rangephage �. This is, to our knowledge, the first example of a quorum-sensing-regulated antiphage defense mechanism. We proposethat this mechanism serves to protect E. coli under conditions ofhigh risk of infection, namely, during growth in high-cell-density,mixed-species environments, where the quorum-sensing signalswould accumulate. Quorum-sensing control of phage susceptibil-ity may be a general phenomenon in microbial communities.

RESULTS� phages accumulate in AHL-treated cultures due to reducedsuperinfection of lysogenic cells. Many known prophages can beinduced to follow the lytic pathway and kill the host cell to releasephage progeny. These prophages have evolved to incorporate sen-sory inputs into the genetic switches that govern this developmen-tal decision. Known induction signals generally provide the pro-phage with information on the metabolic state and stress level ofthe host cell (15–17). As a culture of � lysogens grows, a smallfraction (10�5) of the resident � prophages will induce and lyse thehost cell to release progeny phages into the medium. To assess theeffect of quorum sensing on the interaction between E. coli andphage �, cultures of lysogenic E. coli BW25113 �i434 were grown inthe presence or absence of a cocktail of synthetic AHL autoinducermolecules, and free � phages were enumerated as PFUs on a lawnof �-sensitive bacteria. The cultures were grown at 30°C, whereSdiA has been shown to be most active (18, 19).

Figure 1 shows the concentrations of free � phages in AHL-treated or control lysogenic cultures. In wild-type lysogens, AHLtreatment leads to a 2- to 3-fold increase in free phage levels (bluebars). Thus, quorum-sensing signals do indeed influence phage-bacterium interactions in the classic host-phage pair of E. coli and�. Deletion of the sdiA gene encoding the AHL receptor abolishesthe AHL-mediated increase in free phage (green bars), suggestingthat the AHL effect is mediated by the AHL receptor SdiA. Similarobservations have been reported previously (20). However, themechanism underpinning phage accumulation has not been de-termined and is the focus of the present study.

The concentration of free phages in a culture of lysogens is aresult of two opposing factors, production and loss. First, it de-

pends on the rate of production and release of progeny phages intothe medium by host cell lysis. Second, it depends on the rate of lossof free phages due to superinfection of the remaining cells. Specif-ically, a free � phage can initiate infection of a lysogenic cell byinjecting its DNA into the host cytoplasm, but the superinfectionimmunity system of the resident prophage will prevent expressionof the infecting DNA, leading to loss of the superinfecting phage(21). Hence, the observed increase in free phage in the presence ofquorum-sensing signals could be a result of increased prophageinduction or decreased superinfection of the lysogenic cells orboth. In order to distinguish between these possibilities, we mea-sured the effects of AHLs on prophage induction in bacteria wheresuperinfection is not possible. Specifically, the concentration offree phages in a culture of lysogens immune to superinfection dueto deletion of the � receptor LamB was used as a direct measure ofprophage induction. Since there is no loss of free phage due tosuperinfection, cultures of lamB lysogens show higher concentra-tions of free phages than do wild-type cultures (compare the solidred bar to the solid blue bar in Fig. 1). Importantly, growth in thepresence of AHLs does not stimulate additional phage accumula-tion in the lamB lysogen (red bars). This observation proves thatAHLs do not stimulate induction of the � prophage as suggestedpreviously (20). Rather, the observed AHL effect on free phageconcentration must be due to a decreased loss of free phage, mostlikely due to a reduction in the rate of superinfection of the lyso-genic cells.

A positive regulator of exopolysaccharide synthesis, RcsA (22),was previously reported to be required for the AHL-mediated in-crease in free � phage accumulation (20). Contrary to these re-sults, we do not observe a difference between the wild type and thercsA mutant with respect to the AHL effect on free phage concen-trations (Fig. 1, purple bars).

� phage adsorption rate is reduced when E. coli is grown inthe presence of AHL signals. The results displayed in Fig. 1 showthat superinfection of � lysogens is reduced when the cells aregrown in the presence of AHL autoinducers. We hypothesized

FIG 1 AHL quorum-sensing signals reduce � phage superinfection. Theconcentrations of free phage from lysogenic cultures with or without AHLs areshown for E. coli BW25113 (wild type [WT]), AHL receptor mutant (sdiAmutant), � receptor mutant (lamB mutant), and a mutant of a transcriptionalregulator for exopolysaccharide (rcsA mutant). The cultures were grown to anOD600 of 1.2 in the presence or absence of 5 �M AHLs. Free phage concentra-tion is indicated as PFU per ml of culture. The number of independent culturestested (n) is indicated. Each error bar indicates 1 standard deviation from themean. Reported differences were evaluated using a Student’s t test.

Høyland-Kroghsbo et al.

2 ® mbio.asm.org January/February 2013 Volume 4 Issue 1 e00362-12

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

that AHLs may act to reduce the rate of phage adsorption to thecell surface. To test this hypothesis directly, we subjected wild-type, nonlysogenic E. coli cells grown with or without AHLs toprimary infection by radioactively labeled � phages. After the ad-dition of phages to a cell culture, unadsorbed free phages wereremoved by filtration of an aliquot of the cells every 2 min, and theradioactivity of the adsorbed phages retained on the filters wasquantified (Fig. 2).

Figure 2 shows that AHL treatment of wild-type cells signifi-cantly reduces the rate of � adsorption. The rates of adsorptionwere calculated under the assumption that the decrease in freephage over time follows an exponential decay law (see Fig. S1 inthe supplemental material). The rate of adsorption of � phages toAHL-treated wild-type cells is 2.7 � 10�10 ml min�1 phage�1

cell�1, about half that of untreated cells (4.9 � 10�10 ml min�1

phage�1 cell�1). In contrast, AHL treatment has no effect on the �adsorption rate in the sdiA mutant, demonstrating that the AHLsignals must be transduced through the SdiA receptor to decreasephage adsorption rates. Thus, AHL quorum sensing leads to areduction in the rate of adsorption of phage � to E. coli cells.

AHL-induced downregulation of the � receptor LamB. Themolecular mechanism underlying the reduced infection ratecaused by growth in the presence of AHLs was investigated. Asphage � infects E. coli through the outer membrane maltoporinLamB, the adsorption rate depends on the concentration of LamBreceptors on the cell surface. We therefore investigated whetherAHL autoinducers directly affect LamB protein levels in the outermembrane. Wild-type and sdiA and lamB mutant cells weregrown in the presence or absence of AHLs, and outer membraneproteins were purified and separated by SDS-PAGE (Fig. 3A). Amaltose-induced culture was used as a positive control for identi-fying LamB on the gel. We find that stimulation of wild-type cellsby AHLs lead to a reduction in LamB protein levels by 40%(Fig. 3B). AHL treatment does not affect the LamB levels in thesdiA mutant. This finding clearly shows that AHL signaling

through SdiA leads to a decrease in the levels of LamB in the outermembrane of E. coli BW25113.

AHL signaling dramatically improves E. coli’s chances of sur-viving a virulent phage attack. What is the physiological signifi-cance of the modest reduction in adsorption rate? To evaluate theeffect of quorum sensing on E. coli survival, we exposed E. coli cellsgrown in the presence or absence of AHL signals to a virulentvariant of �, �vir. If the �vir phages were allowed to complete mul-tiple life cycles, they would eventually kill all wild-type cells, andthe only surviving cells would be �-resistant mutants that have lostthe ability to express lamB (23). To address the survival of non-mutant cells, we exposed the cell cultures to a 20-fold excess of �vir

for only 50 min, corresponding to just under one phage genera-tion time under our laboratory conditions. After 50 min, the cat-ions in the medium were chelated to inactivate any free phage andprevent additional rounds of infection. Uninfected cells werecounted as those that grew to form a colony on petri plates thenext day.

Figure 4 shows that the number of E. coli cells surviving �vir

infection is 3 to 4 times higher in cultures that are grown in thepresence of AHLs than in those grown in the absence of AHLs(blue bars). In sdiA mutant cultures, the presence of AHLs do notincrease the chance of surviving the virulent infection, thus dem-onstrating that the AHL effect is entirely dependent on SdiA(green bars). The surviving cells can be divided into two groups,those that are � sensitive but were not infected by a phage for theduration of the experiment and those that carry a mutation thatmakes them genetically resistant to infection by �. To assesswhether the surviving cells in the experiment shown in Fig. 4 aremostly �-resistant mutants or nonmutant cells that have avoided�vir infection, we took advantage of the genetic linkage betweenthe � resistance phenotype and that of the ability to utilize maltoseas a carbon source (24, 25). lamB encodes the maltoporin, and itsexpression is coregulated with the genes required for the uptakeand utilization of maltose. Hence, about 80% of �-resistant mu-

FIG 2 Growth in the presence of AHL signals reduces the rate of phage adsorption. 35S-labeled � phages were added to a shaking culture of wild-type (WT) orAHL receptor mutant (sdiA mutant) cells grown in the presence or absence of 5 �M AHLs. Prior to the addition of phage, chloramphenicol was added to cellsto arrest growth at an OD600 of 1.0 and prevent phage multiplication. At 2-min intervals, an aliquot of the cell-phage mixture was filtered through a 0.45-�m filter,and the radioactivity of the adsorbed phages retained on the filter was measured. The radioactivity of filters subjected to the identical treatment using lamBmutant cells has been subtracted as background, and each sample has been normalized to the total radioactivity of an unfiltered sample aliquot. The number ofindependent cultures tested (n) is indicated. Each error bar indicates 1 standard deviation from the mean. Reported differences were evaluated using a two-wayanalysis of variance (ANOVA). The experiment was repeated on three separate days with similar results. The relative rates of adsorption were calculated as shownin the symbol key to Fig. S1 in the supplemental material and are shown in the figure.

A Quorum-Sensing-Induced Phage Defense Mechanism

January/February 2013 Volume 4 Issue 1 e00362-12 ® mbio.asm.org 3

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

tants have concurrently lost the ability to utilize maltose as a car-bon source (23). These mal mutants can be easily distinguishedfrom mal� cells, as the former form white colonies on maltose-MacConkey indicator agar, while the latter form red colonies.When the surviving cells were plated on maltose-MacConkeyagar, less than 1% of the colonies were white, indicating that thevast majority of surviving cells at the end of the first lytic cycle arenonmutant cells which have simply avoided infection by �vir, pre-sumably due to the display of a reduced number of LamB recep-

tors on their cell surface at the time of phage invasion. The de-creased susceptibility to phage infection is transient, as the vastmajority of daughter cells in colonies arising from the survivorcells are again sensitive to infection by �vir, as determined by cross-streaking those colonies with a high-titer �vir lysate. Thus, AHLquorum sensing through SdiA leads to a severalfold increase in thefraction of an E. coli population that is not resistant, but tran-siently less susceptible, to � phage attack.

AHL quorum-sensing signals also protect E. coli from phage�. AHL signals could also protect E. coli from phages that recog-nize other receptors than LamB. One such example is phage chi(�), which is known to infect E. coli and other enteric bacteriathrough the flagellum (26–28). � phages were added to wild-typeE. coli cells grown in the presence or absence of AHLs, and theproportion of free, nonadsorbed phages was monitored over time.Figure 5 shows that growth in the presence of AHLs significantlyreduces the rate of adsorption of phage � to wild-type cells. Anal-ogous to the � case, the AHL effect depends on the AHL receptorSdiA, as the effect is not observed in sdiA mutant cells.

As we find that AHLs mediate an antiphage defense againstboth phage � and phage �, which have different mechanisms ofinfection, we suggest that quorum sensing may generically activateantiphage strategies in E. coli and potentially other bacterial spe-cies. The generality of this quorum-sensing response is under in-vestigation in our laboratory.

DISCUSSION

Quorum sensing is crucial for the survival and fitness of numerousmicroorganisms, and many variations on the canonical LuxI-LuxR quorum-sensing system have evolved that allow bacteria toengage in communication with their own and other species, en-able quorum quenching and the spread of misinformation, andeven allow crosskingdom signaling with eukaryotic hosts (re-viewed in reference 29). Here, we suggest that bacteria have addi-tionally evolved to use quorum sensing as a means of regulatingtheir interactions with the most abundant biological entities on

FIG 3 AHL signaling induces downregulation of the � receptor LamB. (A) Outer membrane protein preparations were separated by SDS-PAGE and stainedwith Coomassie blue. Outer membrane proteins from wild-type E. coli, AHL receptor mutant (sdiA mutant), and � receptor mutant (lamB mutant) grown to anOD600 of 1.0 with 5 �M AHLs or 0.4% maltose are shown as indicated below the lanes. (B) Quantification of LamB protein. Band intensities of the protein bandwere quantified using ImageJ software and normalized to the intensity of a LamB band that is not affected by AHLs. To enable pooling of the data from differentgels, the intensity of the LamB band in untreated wild-type cells was set at 1 in each gel, and the intensities of the remaining bands relative to that of the untreatedwild-type cells are shown. The number of independent outer membrane protein preparations tested (n) is indicated. Each error bar indicates 1 standard deviationfrom the mean.

FIG 4 AHLs dramatically enhance E. coli’s chances of surviving attack by thelytic phage �vir. Wild-type and sdiA mutant E. coli cells were grown to an OD600

of 1.0 in the presence or absence of 5 �M AHLs. �vir was added at an averagephage input of 20 phages per cell. Fifty minutes after the addition of phage, analiquot of the culture was diluted in M63-salts containing 50 mM sodiumcitrate to inactivate the free phage. The figure shows the number of coloniesformed by cells that survived 50 min in the presence of �vir relative to thenumber of CFU immediately prior to phage addition. The number of inde-pendent cultures tested (n) is indicated. Each error bar indicates 1 standarddeviation from the mean. Reported differences were evaluated using a Stu-dent’s t test.

Høyland-Kroghsbo et al.

4 ® mbio.asm.org January/February 2013 Volume 4 Issue 1 e00362-12

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

Earth, namely, the viruses that prey on them. We propose thatquorum-sensing signals, as a measure of population density, maybe utilized to estimate the local risk of phage infection and toregulate antiviral defense strategies accordingly.

As a first approach to investigate our hypothesis, we evaluatedthe effect of AHL quorum sensing on phage-host interactions inthe classical model system of phage � and E. coli K-12. We foundthat AHL quorum sensing leads to downregulation of the numberof LamB � receptors on the cell surface, which in turn increases thenumber of survivor E. coli cells in a population following a potentphage attack. In addition, we find that AHLs protect E. coli againstinfection by the broad-host-range phage �. Importantly, the ob-served AHL effects depend entirely on the AHL receptor SdiA. Therequirement for SdiA demonstrates that the reduced phage ad-sorption rates in AHL-treated cultures occur as a consequence ofSdiA-AHL-mediated intracellular regulation, not as a side effect ofunexpected physical or chemical changes in the medium causedby the addition of synthetic AHLs.

Although many aspects of E. coli biology are very well under-stood, the role of AHL quorum sensing in E. coli has remainedelusive (30). First, E. coli does not produce AHL molecules andtherefore presumably uses the AHL receptor SdiA exclusively topick up signals released from other AHL-producing bacteria (10).Second, very few genes have reproducibly been shown to be reg-ulated by the AHL-SdiA complex in E. coli K-12 (12–14). One ofthe regulons that are consistently found to be downregulated bySdiA is the flagellar genes (12, 14, 31, 32), which supports ourfinding that AHLs reduce the risk of � infection, as phage � infectsits host through the flagellum. Interestingly, SdiA-mediated tran-scriptional regulation of LamB is supported by results from VanHoudt et al. who found that the AHL variant N-hexanoyl-L-homoserine lactone (C6-HSL) downregulates malT at 30°C inE. coli MG1655, in a plasmid-based promoter trap screen (19). AsMalT is a positive regulator of lamB (33), downregulation of malTwould lead to decreased lamB transcription. The expression ofanother gene in the MalT regulon, malE, was additionally found tobe repressed by plasmid-based expression of sdiA in E. coli K-12grown at 37°C (32). Hence, there are several indications that SdiAmay downregulate � receptor levels via transcriptional repression

of malT, but whether the SdiA-AHL com-plex directly binds and represses the malTpromoter has yet to be elucidated.

The question remains why AHLs thatare not produced by E. coli affect E. coli’sphage defense. Many phages appear tohave extreme host specificity, infectingonly the bacterial strain with which theywere isolated and thus only species-specific, or even strain-specific, quorum-sensing signals would be relevant forregulating phage-host interactions. How-ever, the standard methods for phage iso-lation may have been biased to favorphages of limited host range (35), andbroad-host-range phages are readily iso-lated from environmental samples (34,35). Moreover, phages may adapt to newhost strains at relatively high rates (36).The issue of host specificity in mixed-species microbial communities is still

poorly understood, but it is clear that the subject is more complexthan suggested by the traditional one-phage�one-host view.Thus, in its natural habitats, E. coli may encounter a variety ofbroad-host-range phages. By exploiting quorum-sensing signalsreleased from other bacterial species to reduce its susceptibility tophage infection, E. coli may be able to protect itself against thesephages.

The evolutionary dynamics between phage and host has beendescribed as an arms race, where antagonistic interactions be-tween host and parasite result in ever increasing phage infectivityand bacterial resistance. However, increased bacterial resistancecomes at a price. Hall et al. reported that arms race evolutionresults in decreased relative fitness, and that over time, this dy-namics is replaced by fluctuating selection, potentially due to costsof generalism and mutational limitations (5). Gómez and Buck-ling (4) found that in a natural microbial environment,bacterium-phage coevolution is characterized by fluctuating se-lection and thus generates bacteria that are more resistant to pres-ent phages than to past and future viruses. They suggest that theevolution of resistance toward contemporary phages is favored, asthis is less costly in terms of growth rate than general resistancemechanisms (4). In keeping with this theory, we argue that quo-rum sensing allows bacteria to spare the cost of maintaining aconstantly elevated phage defense. Thus, a quorum-sensing-regulated phage defense mechanism could potentially play a keyrole for bacterium-phage coevolution in natural environments.

Phage-host systems show unusual prey-predator dynamics be-cause the consumption of one prey leads to the generation of tensto hundreds of new predators. Given this, it is remarkable thatvirulent phages do not always drive their host population to ex-tinction. To explain phage-host coexistence, various mechanismshave been suggested that would generate a minority group of hostcells that are protected from infection, either due to spatial sepa-ration (37), decreased lysis in stationary-phase cells (38), or sto-chastic fluctuations in phage receptor expression (39). We pro-pose that by subjecting genes encoding phage receptors (lamB andflagellar genes) to quorum-sensing control, E. coli has evolved theability to increase the subpopulation of phage-tolerant receptor-free cells at times where the risk of phage infection is elevated,

FIG 5 AHL signaling protects E. coli from � phage adsorption. Wild-type E. coli was grown to an OD600

of 0.75 in the presence or absence of 5 �M AHLs. To arrest growth and prevent phage multiplication,chloramphenicol was added prior to the addition of phage �. The numbers of free (nonadsorbed) �phages were measured as PFU on a lawn of sensitive, motile bacteria. The number of independentcultures tested (n) is indicated. Each error bar represents 1 standard deviation from the mean. Reporteddifferences were evaluated using a two-way analysis of variance (ANOVA).

A Quorum-Sensing-Induced Phage Defense Mechanism

January/February 2013 Volume 4 Issue 1 e00362-12 ® mbio.asm.org 5

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

despite the cost associated with temporary loss of the primaryfunction of these phage receptors (uptake of maltodextrins andmotility). Quorum-sensing-mediated downregulation of phagesusceptibility may well turn out to be an important factor in un-derstanding phage-host coexistence. The phenomenon bears astriking similarity to the formation of population heterogeneity inother contexts, such as the generation of competent subpopula-tions, sporulating subpopulations, or subpopulations that persistdespite antibiotic treatment. Notably, quorum sensing has beenfound to upregulate the generation of all of these minority groups(40–43).

Phage-host interactions are understood better in the classicalmodel system of bacteriophage � and E. coli K-12 than in any othersystem. However, despite its paradigmatic status, new knowledgecontinues to be revealed as exemplified here by a novel and im-portant role for quorum sensing in regulating phage susceptibil-ity. Given the incompleteness of our understanding of this rela-tively simple system, one can only speculate how far we must befrom an in-depth understanding of phage-host interactions incomplex natural niches, such as the human gut. The launch of thehuman microbiome project marks a renaissance in studies of thecontributions of the human bacterial flora to well-being andpromises to deliver new treatments for lifestyle diseases throughengineering alterations in the gut flora (reviewed in references 44,45, and 46). However, for these efforts to succeed, a comprehen-sive understanding of the factors that naturally shape the size,composition, structure, and development of microbial communi-ties is required. On the basis of our finding in the E. coli-� modelsystem, we hypothesize that an important, but so far overlooked,determinant of phage-host dynamics could be bacterial cell-cellcommunication.

MATERIALS AND METHODSBacterial strains and bacteriophages. The bacterial strains used in thisstudy are all derivatives of Escherichia coli K-12. The bacterial strains,phages, and plasmids used in this study are listed in Table S1 in the sup-plemental material.

Strain construction. A single �i434 lysogen of E. coli BW25113, iden-tified by the PCR assay described by Powell et al. (47), gave rise to strainJMØ3. Gene replacement of the open reading frames (ORFs) of lamB,sdiA, and rcsA with an antibiotic resistance cassette was performed by themethod of Datsenko and Wanner (48). Antibiotic-resistant transfor-mants were screened by PCR with primers flanking the ORF to identifytransformants with gene replacements of the expected size. The PCRprimers used to amplify the antibiotic resistance cassette are listed in Ta-ble S2 in the supplemental material.

AHL preparation. A mixture of 6 AHLs (N-(butyl, heptanoyl,hexanoyl, ketocaproyl, octanoyl, and tetradecanoyl)-DL-homoserine lac-tones) were dissolved in ethyl acetate acidified by 0.1% acetic acid (EA)and stored at �20°C. Prior to each experiment, the AHL mix or the equiv-alent volume of EA alone was added to glass culture tubes to a final AHLconcentration of 5 �M and incubated with shaking at room temperatureuntil completely dry as in reference 20.

Bacterial growth conditions. For all experiments, E. coli BW25113and its derivatives were grown from single colonies in TB medium (15 gtryptone and 5 g NaCl per liter) at 30°C, shaking at 220 rpm. Exponen-tially growing cultures were diluted a minimum of 1,000-fold into TBmedium in glass tubes coated with the AHL mixture or control tubestreated with EA as described above and allowed to reach the desired celldensity. Cell densities were measured by determining the optical densityat 600 nm (OD600) on an Ultraspec 2100 Pro (Amersham Biosciences).One OD600 unit corresponds to a cell density of 1.1 � 109 CFU/ml.

Measurement of free phage in lysogenic cultures. E. coli BW25113single lysogens of �i434 and otherwise isogenic lamB, sdiA, and rcsA mu-tant strains were grown to an OD600 of 1.2, and cells were pelleted bycentrifugation for 5 min at 7,500 � g. A drop of chloroform was added tothe supernatants. The concentration of free phages in the supernatant wasquantified as the number of PFU on a lawn of �-sensitive, maltose-induced E. coli MG1655 as described previously (49). Three indepen-dently constructed rcsA mutants were included in the study to verify theAHL-dependent phenotype observed in this mutant.

35S labeling of bacteriophage �. Single colonies of E. coli S2775 weregrown for 5 to 6 h at 30°C in sulfate-free M63 (30.00 g KH2PO4, 91.71 gK2HPO4·3H2O, 16.19 g NH4Cl, 6.54 mg FeCl2·4H2O, 9.68 mg Na3citrateper liter) supplemented with Na2SO4 to a final concentration of 276 �M,followed by a 15-min incubation at 42°C to induce the temperature-sensitive prophage. To remove sulfate, cells were washed twice in sulfate-free M63, and the growth medium was supplemented with 500 �Ci 35S-labeled methionine and 2 �g unlabeled methionine. The cultures wereincubated with shaking for 2 h at 42°C to allow phage proliferation. Then,the lysate was cleared by the addition of a few drops of chloroform fol-lowed by centrifugation for 10 min at 8,000 � g to remove cell debris. Thesupernatant was transferred to a fresh tube and subjected to cesium chlo-ride banding to purify the phages as previously described (50).

Adsorption assay of 35S-labeled phages. Wild-type and lamB andsdiA mutant cells were grown in the presence or absence of AHLs to anOD600 of 1.0, and chloramphenicol was added to a final concentration of0.05 mg/ml to prevent further growth. To begin the assay, 1.6 �108 PFU/ml of 35S-labeled phages were added to the shaking cell culture,and aliquots were filtered 2, 4, 6, 8, and 10 min after the addition of phage.Filtration through a 0.45-�m membrane filter (Durapore membrane fil-ter [catalog no. HVLP02500; Millipore]) served to retain bacterial cellsand any adsorbed phages on the filter, while allowing free phages to passthrough. The radioactivity retained on the filters was quantified using aPerkinElmer Wallac 1414 liquid scintillation counter, counting each filterfor 20 min. For each cell culture, an unfiltered sample aliquot was countedto quantify the total radioactivity of the phage-cell mixture. Adsorptionvalues in Fig. 2 are calculated as follows: (counts on filter minus the aver-age counts retained on filters with lamB mutant cells)/(total counts ofsample aliquot).

Survival assay. Cultures of wild-type and sdiA mutant cells weregrown in the presence or absence of AHLs to an OD600 of 1.0. �vir wasadded at an average phage input of 20 phages per cell. Samples werecollected immediately before and 50 min after phage addition and dilutedinto ice-cold M63-salts containing 50 mM sodium citrate to inactivate anyfree phages. The dilutions were plated on 0.4% maltose-MacConkeyplates containing 5 mM sodium citrate and incubated at 30°C overnight.Surviving cells were counted as red and white colonies the following day.Two hundred colonies were cross-streaked against �vir to determinewhether survival was due to transient phage tolerance (�vir-sensitive col-ony) or inherited phage resistance (�vir-resistant colony).

Outer membrane protein preparations and SDS-PAGE. Cultureswere grown in the presence or absence of AHLs to an OD600 of 1.0, har-vested, and chilled on ice. All steps were carried out cold, except whennoted otherwise. Cell pellets were collected by centrifugation for 5 min at6,000 � g at 4°C. Pellets were washed in 1 ml of 50 mM Tris HCl (pH 8)and resuspended in 20% sucrose and 100 mM Tris HCl (pH 8). Lysozymeand EDTA were added to final concentrations of 0.4 mg/ml and 10 mM,respectively. Samples were incubated at 4°C overnight. The next day, thesamples were overlaid by an equal volume of 2% Triton X-100, 10 mMMgCl2, and 100 mM Tris HCl (pH 8) and sonicated until clear. Outermembrane proteins were collected by centrifugation for 22.5 min at16,000 � g at 4°C. Membrane particles were washed once in 50 mM TrisHCl (pH 8), 5 mM MgCl2, and 1% Triton X-100 and spun down at 16,000� g and 4°C for 15 min. Membrane particles were washed twice with50 mM Tris HCl (pH 8) and 5 mM MgCl2 to remove Triton X-100. Pelletswere resuspended at room temperature in 100 mM Tris HCl (pH 8) and

Høyland-Kroghsbo et al.

6 ® mbio.asm.org January/February 2013 Volume 4 Issue 1 e00362-12

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

2% SDS. Protein concentrations were determined using a NanoDrop1000 (Thermo Scientific). This protocol was adapted from references 51and 52. Proteins were separated on a 10% SDS-polyacrylamide gel andstained with Coomassie blue. Protein band intensities were quantifiedusing ImageJ and normalized to a band that was constantly expressedregardless of the presence of AHLs.

� phage assay. Wild-type cells were grown in the presence or absenceof AHLs to an OD600 of 0.75, and chloramphenicol was added to a finalconcentration of 0.05 mg/ml to prevent further growth. � phage lysate wasadded to a final concentration of 1.6 � 108 PFU/ml to begin the assay. Ateach time point, a culture aliquot was diluted into ice cold TMG (0.121%Tris base, 0.12% MgSO4, 0.01% gelatin [pH 7.4]) to prevent further ad-sorption, and cells were removed by centrifugation for 5 min at 7,500 � g.Chloroform was added to the supernatant, and free, nonadsorbed �phages were quantified by plating on motile, �-sensitive cells, and count-ing PFUs. Sensitive chloramphenicol-resistant cells were prepared bygrowing E. coli KX1440 cells to an OD600 of 0.75. � phage dilutions weremixed with sensitive cells, incubated 10 min at 37°C, plated with TB mo-tility agar (0.4% agar) on a solid TB plate, and incubated at 30°C over-night.

Statistical analyses. Reported differences were evaluated using a Stu-dent’s t test for individual measurements (Fig. 1 and 4) or a two-wayanalysis of variance (ANOVA) for data containing repeated measure-ments of the same cultures (Fig. 2 and 5). The analyses were carried outwith GraphPad Prism 5 software.

SUPPLEMENTAL MATERIALSupplemental material for this article may be found at http://mbio.asm.org/lookup/suppl/doi:10.1128/mBio.00362-12/-/DCSupplemental.

Figure S1, TIF file, 0.1 MB.Table S1, PDF file, 0.4 MB.Table S2, PDF file, 0.2 MB.

ACKNOWLEDGMENTS

This work was supported in part by a research grant to S.L.S. from theNovo Nordisk Foundation. N.M.H.-K. was supported by the Danish Na-tional Research Foundation through the Center for Models of Life. We aregrateful to Stanley Brown and Steen Pedersen for their help with generat-ing the radioactive phage lysate. Special thanks to Linda Hove Chris-tensen, Julia Madsen-Østerbye, and Samantha Steen for their contribu-tions in the project’s earliest stages. We thank Stanley Brown, SaeedTavazoie, and Karina Xavier for bacterial strains and phages.

REFERENCES1. Ng WL, Bassler BL. 2009. Bacterial quorum-sensing network architec-

tures. Annu. Rev. Genet. 43:197–222.2. Brüssow H, Hendrix RW. 2002. Phage genomics: small is beautiful. Cell

108:13–16.3. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage resistance mech-

anisms. Nat. Rev. Microbiol. 8:317–327.4. Gómez P, Buckling A. 2011. Bacteria-phage antagonistic coevolution in

soil. Science 332:106 –109.5. Hall AR, Scanlan PD, Morgan AD, Buckling A. 2011. Host-parasite

coevolutionary arms races give way to fluctuating selection. Ecol. Lett.14:635– 642.

6. Lederberg EM. 1950. Lysogenicity in Escherichia coli strain K-12. Microb.Genet. Bull. 1:5– 8.

7. Gottesman ME, Weisberg RA. 2004. Little lambda, who made thee?Microbiol. Mol. Biol. Rev. 68:796 – 813.

8. Sitnikov DM, Schineller JB, Baldwin TO. 1996. Control of cell division inEscherichia coli: regulation of transcription of ftsQA involves both rpoSand SdiA-mediated autoinduction. Proc. Natl. Acad. Sci. U. S. A. 93:336 –341.

9. Wang XD, de Boer PA, Rothfield LI. 1991. A factor that positivelyregulates cell division by activating transcription of the major cluster ofessential cell division genes of Escherichia coli. EMBO J. 10:3363–3372.

10. Soares JA, Ahmer BM. 2011. Detection of acyl-homoserine lactones byEscherichia and Salmonella. Curr. Opin. Microbiol. 14:188 –193.

11. Yao Y, Martinez-Yamout MA, Dickerson TJ, Brogan AP, Wright PE,Dyson HJ. 2006. Structure of the Escherichia coli quorum sensing proteinSdiA: activation of the folding switch by acyl homoserine lactones. J. Mol.Biol. 355:262–273.

12. Dyszel JL, Soares JA, Swearingen MC, Lindsay A, Smith JN, Ahmer BM.2010. E. coli K-12 and EHEC genes regulated by SdiA. PLoS One 5:e8946.http://dx.doi.org/10.1371/journal.pone.0008946.

13. Hughes DT, Terekhova DA, Liou L, Hovde CJ, Sahl JW, Patankar AV,Gonzalez JE, Edrington TS, Rasko DA, Sperandio V. 2010. Chemicalsensing in mammalian host-bacterial commensal associations. Proc. Natl.Acad. Sci. U. S. A. 107:9831–9836.

14. Kanamaru K, Tatsuno I, Tobe T, Sasakawa C. 2000. SdiA, an Escherichiacoli homologue of quorum-sensing regulators, controls the expression ofvirulence factors in enterohaemorrhagic Escherichia coli 0157:H7. Mol.Microbiol. 38:805– 816.

15. Little JW, Mount DW. 1982. The SOS regulatory system of Escherichiacoli. Cell 29:11–22.

16. Melechen NE, Go G. 1980. Induction of lambdoid prophages by aminoacid deprivation: differential inducibility; role of recA. Mol. Gen. Genet.180:147–155.

17. Ptashne M. 1986. A genetic switch: gene control and phage lambda. CellPress, Cambridge, MA.

18. Lee J, Zhang XS, Hegde M, Bentley WE, Jayaraman A, Wood TK. 2008.Indole cell signaling occurs primarily at low temperatures in Escherichiacoli. ISME J. 2:1007–1023.

19. Van Houdt R, Aertsen A, Moons P, Vanoirbeek K, Michiels CW. 2006.N-Acyl-L-homoserine lactone signal interception by Escherichia coli.FEMS Microbiol. Lett. 256:83– 89.

20. Ghosh D, Roy K, Williamson KE, Srinivasiah S, Wommack KE, Ra-dosevich M. 2009. Acyl-homoserine lactones can induce virus productionin lysogenic bacteria: an alternative paradigm for prophage induction.Appl. Environ. Microbiol. 75:7142–7152.

21. Lwoff A. 1953. Lysogeny. Bacteriol. Rev. 17:269 –337.22. Majdalani N, Gottesman S. 2005. The Rcs phosphorelay: a complex signal

transduction system. Annu. Rev. Microbiol. 59:379 – 405.23. Thirion JP, Hofnung M. 1972. On some genetic aspects of phage lambda

resistance in E. coli K12. Genetics 71:207–216.24. Randall-Hazelbauer L, Schwartz M. 1973. Isolation of the bacteriophage

lambda receptor from Escherichia coli. J. Bacteriol. 116:1436 –1446.25. Szmelcman S, Hofnung M. 1975. Maltose transport in Escherichia coli

K-12: involvement of the bacteriophage lambda receptor. J. Bacteriol. 124:112–118.

26. Iino T, Mitani M. 1967. Infection of Serratia marcescens by bacteriophagechi. J. Virol. 1:445– 447.

27. Meynell EW. 1961. A phage, Phi chi, which attacks motile Bacteria. J. Gen.Microbiol. 25:253–290.

28. Sertic V, Boulgakov NA. 1936. Specific bacteriophages for flagellatedbacterial varieties. C. R. Soc. Biol. 123:887– 888.

29. Bassler BL, Losick R. 2006. Bacterially speaking. Cell 125:237–246.30. Fuqua C, Greenberg EP. 2002. Listening in on bacteria: acyl-homoserine

lactone signalling. Nat. Rev. Mol. Cell Biol. 3:685– 695.31. Sharma VK, Bearson SM, Bearson BL. 2010. Evaluation of the effects of

sdiA, a luxR homologue, on adherence and motility of Escherichia coli0157:H7. Microbiology 156:1303–1312.

32. Wei Y, Lee JM, Smulski DR, LaRossa RA. 2001. Global impact of sdiAamplification revealed by comprehensive gene expression profiling ofEscherichia coli. J. Bacteriol. 183:2265–2272.

33. Boos W, Shuman H. 1998. Maltose/maltodextrin system of Escherichiacoli: transport, metabolism, and regulation. Microbiol. Mol. Biol. Rev.62:204 –229.

34. Comeau AM, Chan AM, Suttle CA. 2006. Genetic richness of vibrio-phages isolated in a coastal environment. Environ. Microbiol.8:1164 –1176.

35. Jensen EC, Schrader HS, Rieland B, Thompson TL, Lee KW, NickersonKW, Kokjohn TA. 1998. Prevalence of broad-host-range lytic bacterio-phages of Sphaerotilus natans, Escherichia coli, and Pseudomonas aerugi-nosa. Appl. Environ. Microbiol. 64:575–580.

36. Letarov A, Kulikov E. 2009. The bacteriophages in human- and animalbody-associated microbial communities. J. Appl. Microbiol. 107:1–13.

37. Heilmann S, Sneppen K, Krishna S. 2010. Sustainability of virulence in aphage-bacterial ecosystem. J. Virol. 84:3016 –3022.

38. Weitz J, Dushoff J. 2008. Alternative stable states in host-phage dynam-ics. Theor. Ecol. 1:13–19.

A Quorum-Sensing-Induced Phage Defense Mechanism

January/February 2013 Volume 4 Issue 1 e00362-12 ® mbio.asm.org 7

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

39. Chapman-McQuiston E, Wu XL. 2008. Stochastic receptor expressionallows sensitive bacteria to evade phage attack. Part I. Experiments. Bio-phys. J. 94:4525– 4536.

40. Grossman AD, Losick R. 1988. Extracellular control of spore formationin Bacillus subtilis. Proc. Natl. Acad. Sci. U. S. A. 85:4369 – 4373.

41. Leung V, Levesque CM. 2012. A stress-inducible quorum-sensing pep-tide mediates the formation of noninherited multidrug tolerance. J. Bac-teriol. 194:2265–2274.

42. Magnuson R, Solomon J, Grossman AD. 1994. Biochemical and geneticcharacterization of a competence pheromone from B. subtilis. Cell 77:207–216.

43. Solomon JM, Lazazzera BA, Grossman AD. 1996. Purification and char-acterization of an extracellular peptide factor that affects two differentdevelopmental pathways in Bacillus subtilis. Genes Dev. 10:2014 –2024.

44. Hattori M, Taylor TD. 2009. The human intestinal microbiome: a newfrontier of human biology. DNA Res. 16:1–12.

45. Sekirov I, Russell SL, Antunes LC, Finlay BB. 2010. Gut microbiota inhealth and disease. Physiol. Rev. 90:859 –904.

46. Tilg H, Kaser A. 2011. Gut microbiome, obesity, and metabolic dysfunc-tion. J. Clin. Invest. 121:2126 –2132.

47. Powell BS, Rivas MP, Court DL, Nakamura Y, Turnbough CL, Jr.. 1994.Rapid confirmation of single copy lambda prophage integration by PCR.Nucleic Acids Res. 22:5765–5766.

48. Datsenko KA, Wanner BL. 2000. One-step inactivation of chromosomalgenes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci.U. S. A. 97:6640 – 6645.

49. Silhavy TJ, Berman ML, Enquist LW. 1984. Experiments with genefusions. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

50. Miller JH. 1972. Experiments in molecular genetics. Cold Spring HarborLaboratory, Cold Spring Harbor, NY.

51. Brown S. 1992. Engineered iron oxide-adhesion mutants of the Esche-richia coli phage lambda receptor. Proc. Natl. Acad. Sci. U. S. A. 89:8651– 8655.

52. Heine HG, Francis G, Lee KS, Ferenci T. 1988. Genetic analysis ofsequences in maltoporin that contribute to binding domains and porestructure. J. Bacteriol. 170:1730 –1738.

Høyland-Kroghsbo et al.

8 ® mbio.asm.org January/February 2013 Volume 4 Issue 1 e00362-12

m

bio.asm.org

on April 13, 2014 - P

ublished by m

bio.asm.org

Dow

nloaded from

Genotype Source or reference

Strains

BW25113 Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), rph-1, Δ(rhaD-

rhaB)568, hsdR514

(Datsenko and Wanner 2000)

MG1655 F-, rph-1 (52)

JMØ3 BW25113 λi434 single lysogen This study

JMØ11 JMØ3 rcsA::cat #1 This study

SLS4146 JMØ3 rcsA::cat #2 This study

SLS4147 JMØ3 rcsA::cat #3 This study

SLS4141 BW25113 lamB::cat This study

SLS4143 JMØ3 lamB::cat This study

NMHK8 BW25113 sdiA::cat This study

NL1A JMØ3 sdiA::cat This study

S2775 λcI857 temperature inducible single lysogen (53)

KX1440 MG1655 lsrK::cat K. Xavier lab collection,

constructed according to (54)

Phages

λi434 λ hybrid containing the immunity region of coliphage 434 (55).

Gift from Stanley Brown,

University of Copenhagen,

Denmark.

λvir virulent λ mutant, which is insensitive to the CI repressor. (56).

Gift from Stanley Brown,

University of Copenhagen,

Denmark.

χ Broad host range phage Gift from Saeed Tawazoie,

Columbia University, USA. (27)

Plasmids

pKD46 Expressing the λ Red recombinase system (48)

pKD3 Template for chloramphenicol resistance cassette (48)

Target

Gene

Forward Primer 5’-3’ Reverse Primer 5’-3’

rcsA

CAGAGTAATCTTATTCAGCCTGACTGGTGGGAAACCA

CCAGTCAGATTGTGCATATGAATATCCTCCTTA

GAAATATTCAGGTAAGGGGAATTATCGTTACGCATTGAG

TGAGGGTATGCCGTGTAGGCTGGAGCTGCTTC

lamB

GCCTGTCACAGGTGATGTGAAAAAAGAAAAGCAATG

ACTCAGGAGATAGATGTGTAGGCTGGAGCTGCTTCG

GTTGGTTGCCGAATGCGGCGTAAACGCCTTATCCGGCCC

AGGTTTTGCTACATATGAATATCCTCCTTA

sdiA CGTGCCTTTCAGCCGGTTTTTGCATCTGGCACGCAGG

ACAGAAAAGAGACATATGAATATCCTCCTTA

GTTTGAATTATCATTATAAATGATACTCACTCTCAGGGGC

GTTGCGGTTTACTGTGTAGGCTGGAGCTGCTTC

41

Supplemental figure S1, related to figure 2.

42

4. Manuscript A: AHL quorum sensing upregulates cnu, a member of the Hha-family of H-NS modulatory proteins in E. coli

43

AHL quorum sensing upregulates cnu, a member of the Hha-family of H-NS modulatory proteins in E. coli Nina Molin Høyland-Kroghsbo and Sine Lo Svenningsen Institute of Biology, University of Copenhagen, Copenhagen, Denmark

Abstract

Quorum sensing is an important bacterial cell-cell communication process that allows bacteria to coordinate group-based behavior in response to the population density. This allows bacteria to unite to perform tasks that single cells would not achieve, such as virulence, biofilm formation, and bioluminescence. Curiously, E. coli harbors a receptor protein, SdiA, which detects acyl homoserine lactone (AHL) quorum sensing signals, but does not itself contain an enzyme capable of synthesizing AHL. SdiA is therefore believed to allow E. coli to respond to quorum sensing signals produced by other bacterial species in its surroundings. H-NS is a highly abundant global transcriptional silencer, that preferentially binds dsDNA but also associates directly with a limited number of mRNAs with inherently poor Shine-Dalgarno sequences, to exert positive posttranscriptional regulation. Specific environmental signal relieve repression of selected H-NS targets as a result of a complex network of factors that can affect the expression or activity of this important regulator. H-NS form heteromeric complexes with its close paralogue StpA and the homologues Hha and Cnu. We find that AHL quorum sensing signaling enhances cnu expression in E. coli, and that enhanced cnu expression results in reduced translation of the H-NS-activated malT mRNA. We propose that enhanced Cnu levels changes the composition of H-NS heteromeric complexes and thus affects H-NS target affinity. The SdiA-AHL complex is the first reported activator of cnu expression, and our finding thus adds AHL quorum sensing signals to the list of environmental signals that may modulate H-NS activity.

Introduction

The nucleoid-associated DNA-binding protein H-NS is a highly abundant global transcriptional silencer, present in about 20,000 copies pr. cell in gram-negative bacteria (reviewed in (Dorman, 2014)). Although a few cases of H-NS regulation of RNA targets have been described (Brescia et al., 2004; Park et al., 2010), it preferentially binds intrinsically curved AT-rich DNA sequences, including horizontally acquired virulence genes in E. coli and S. Typhimurium. It is believed that one role of H-NS is to mediate repression of horizontally acquired genes to prevent their inappropriate expression, and that modulators of H-NS expression allows for the bacterium to express these under specific circumstances (Dorman, 2014; Stoebel et al., 2008). Known modulators of H-NS activity include StpA, which is a close paralogue of H-NS. Both proteins are composed of an N-terminal oligomerization domain attached via a flexible linker to a C-terminal nucleic acid binding domain. H-NS and the approximately 7-fold less abundant StpA (Wang et al., 2012) form homo- and heterodimers and StpA can partially compensate for the loss of H-NS (Stoebel et al., 2008). In addition, Cnu (also known as YdgT) and Hha belong to a family of

44

small proteins that are homologous to the oligomerization domain of H-NS, but lack a DNA-binding domain. Cnu and Hha are the only two chromosomally encoded members of the Hha family in E. coli (Madrid et al., 2007). Cnu and Hha are approximately 130- and 5.000-fold less abundant than H-NS, respectively (Wang et al., 2012). Paytubi et al. investigated the interactions of H-NS, StpA, Hha, and Cnu. They found that Cnu forms complexes with both H-NS and StpA in vitro and that Hha forms a complex with StpA. They did not find an interaction between Cnu and Hha and proposed that each of these can form a heterodimeric complex with either H-NS or StpA (Paytubi et al., 2004). They also found that a hha mutant had increased cnu expression and that cnu could partly compensate for the loss of hha. They further hypothesized that Hha and Cnu proteins have evolved to become adaptable modules that by binding either H-NS, StpA, or both, can change their DNA-binding properties and thus provide an additional level of gene regulation (Paytubi et al., 2004). In order to survive, bacteria continually have to sense and respond to changes in their environment. One important group of signals that bacteria may encounter is autoinducers (AIs). AIs are small diffusible molecules that are continually produced by bacteria and emitted into the environment. As a population of AI-producing bacteria increases in density, AIs accumulate, and when a certain threshold AI concentration is detected, the bacterial population simultaneously turns on a number of group-behavior genes, thus allowing the bacteria to synchronize behaviors across the population. This process, known as quorum sensing (QS), is a key regulator of important bacterial functions, including bioluminescence, biofilm formation, and virulence (reviewed in (Ng and Bassler, 2009)). Both inter- and intra-species AIs exist. Gram-negative bacteria typically produce and detect a class of intra-species AIs called N-acyl-L-homoserine lactones (AHLs). The length and modifications on the acyl side chain is typically specific to a particular bacterial species. Curiously, E. coli, S. Typhimurium, as well as other enteric bacteria lack AHL synthase enzymes, and are thus incapable of AHL production. Nonetheless, via the orphan AHL receptor SdiA, E. coli and S. typhimurium are able to detect the presence of other Gram-negative species, which produce AHLs (reviewed in (Patankar and Gonzalez, 2009)). SdiA encodes a LuxR-type transcription factor, which is activated by a broad range of AHL molecules, but is especially responsive to the AHL variants oxo-C6-HSL and oxo-C8-HSL (Michael et al., 2001). An sdiA mutant of enterohemorrhagic E. coli (EHEC) O157:H7 is incapable of colonizing cattle (Hughes et al., 2010), but despite numerous studies of the SdiA-regulon, only a few targets of the AHL-SdiA complex have been conclusively identified. If considering only the studies of physiologically relevant concentrations of SdiA expressed from its chromosomal location, SdiA inhibits flagellar synthesis in E. coli K-12 and EHEC, and virulence genes in the locus of enterocyte effacement (LEE) pathogenicity island of EHEC (Dyszel et al., 2010; Hughes et al., 2010; Sharma et al., 2010) and positively regulates glutamate-dependent acid resistance genes (Dyszel et al., 2010; Hughes et al., 2010; Van Houdt et al., 2006). We previously reported that AHL-SdiA signaling controls a bacteriophage defense mechanism (Hoyland-Kroghsbo et al., 2013). Specifically, we showed that upon AHL QS signaling, E. coli downregulates the amount of LamB maltoporin in its outer membrane, which is utilized as a

45

receptor for infection by the λ phage (Randall-Hazelbauer and Schwartz, 1973). Thereby, AHL-sensing E. coli effectively reduce the rate of phage adsorption, which in turn leads to a fourfold increase in the fraction of surviving bacteria upon a potent λ phage attack. We also found that AHL protects E. coli from infection by the broad-host-range χ-phage, known to infect E. coli and other enteric bacteria through the flagellum (Iino and Mitani, 1967; Meynell, 1961; Sertic and Boulgakov, 1936). As bacteriophages require a bacterial host to propagate, it follows that a bacterium in a high-cell-density bacterial population is at higher risk of bacteriophage infection than a bacterium in a sparsely populated environment. Hence, we proposed that E. coli uses high QS signals as a cue of high risk of bacteriophage infection. In this study, we sought to unravel the regulatory pathway underlying the AHL-mediated downregulation of LamB in E. coli. LamB functions to transport maltodextrins across the outer membrane, and is encoded in the malB region of the E. coli chromosome. lamB expression is activated by the MalT protein, which activates the entire maltose regulon, consisting of the malA and malB regions (reviewed in (Boos and Shuman, 1998)). Interestingly, a small, but reproducible effect of AHL on malT was previously reported in a promoter-trap screen (Van Houdt et al., 2006). Furthermore, another gene in the MalT regulon, malE, was found to be repressed by plasmid-based overexpression of sdiA in E. coli K-12 grown at 37°C (Wei et al., 2001). Although overexpression of SdiA has been found to affect genes that do not appear to be regulated by physiological levels of SdiA (Dyszel et al., 2010), the data suggests that SdiA may down-regulate LamB λ receptor levels via repression of the positive regulator malT. Although H-NS is primarily known for its role as a DNA-binding protein, a few cases of direct translational activation by H-NS binding to specific mRNAs have been reported. Of these, H-NS regulation of malT is the best described. First Johansson et al. found that H-NS, and to a lesser extend its paralogue StpA, are required for full expression of MalT protein and lamB and malE mRNA (Johansson et al., 1998). Park et al. further dissected the underlying mechanism and elegantly showed that H-NS binds malT mRNA directly in the -35 region relative to the start codon and this in turn repositions the 30S preinitiation complex into a more favorable position and thus enhances translation. WT malT has a poor Shine-Dalgarno (SD) sequence and Park et al. showed that a malT transcript carrying an optimized SD region was highly translated independently of H-NS. Based on the hypothesis that H-NS may enhance translation of other mRNAs with suboptimal SD sequences, they discovered several transcripts regulated by the same mechanism (Park et al., 2010). Notably, they found that H-NS enhances translation of lrhA, encoding a repressor of flhDC, the master regulator of flagella synthesis (Lehnen et al., 2002), thus linking H-NS to regulators of both LamB and flagella, the receptors for the λ and χ phages respectively, both of which are repressed by AHL (Hoyland-Kroghsbo et al., 2013). Except from a previous observation that cnu expression was increased in an hha mutant background (Paytubi et al., 2004), to the best of our knowledge, we here describe SdiA as the first regulator of cnu expression.

46

Results

AHL downregulates the mal operon proteins LamB and MalE In order to test the effect of AHL QS signaling on mal operon expression, we evaluated the effect of AHL on the protein levels of LamB and MalE, encoded in the malA and malB chromosomal regions respectively (Boos and Shuman, 1998). E. coli MG1655 and an isogenic AHL-receptor sdiA mutant were grown in the presence of either oxo-C6-AHL or a solvent control. The western blot in Figure 1 shows that AHL QS signaling leads to a down regulation of both LamB and MalE proteins. This finding suggests that AHL-mediated repression is not specific to LamB, but common to the proteins encoded in the MalT regulon. Importantly, AHL-mediated repression of LamB and MalE is dependent on the AHL-receptor SdiA.

Fig. 1. The effect of AHL on LamB and MalE expression. The relative LamB and MalE protein expression is shown for WT E. coli MG1655 and an isogenic AHL receptor mutant, sdiA. The cultures were grown to an OD600 of 0.75 in the presence of 5 µM oxo-C6-AHL or a solvent control. Whole-cell lysates of 0.1 OD unit cells were subjected to SDS-PAGE and western blotting using anti-LamB and anti-MalE antibodies. The data is representative of at least 3 independent experiments.

AHL regulates malT posttranscriptionally

To determine the level at which AHL regulates malT, we assayed a translational lamB∆60’-‘lacZ reporter (Emr and Silhavy, 1980), a transcriptional malT’-lacZ+ reporter, and a translational malT’-‘lacZ reporter (Raibaud et al., 1991) for their β-galactosidase activity in response to AHL. The fusions were integrated at the lamB- or malT locus on the chromosome of E. coli MG1655 respectively. The lamB∆60’-‘lacZ reporter is repressed by AHL, consistent with the AHL-mediated repression of LamB we detected by western blot. The transcriptional malT’-lacZ+ fusion is not under control by AHL QS signaling (Fig. 2). The translational malT’-‘lacZ fusion, however, is inhibited by AHL and this regulation is dependent on SdiA (fig. 2, green bars). These results indicate that AHL QS signaling inhibits lamB expression by posttranscriptional repression of its positive regulator malT.

47

Fig. 2. Effect of AHL on fusions of lamB and malT to lacZ. E. coli MG1655 (blue bars) or an isogenic sdiA mutant (green bars) harboring either a translational lamB∆60’-‘lacZ fusion, a transcriptional malT’-lacZ+ fusion or a translational malT’-‘lacZ fusion were grown at 30°C in TB in with 5 µM oxo-C6-AHL or a solvent control. Relative β-galactosidase activities were assayed at an OD600 of 1. Error bars show standard deviations of three independent biological replicates. The data is representative of at least 3 independent experiments. The MalT translational fusion protein contains only the first 9 amino acids of MalT (Raibaud et al., 1991), which makes posttranslational regulation by AHL unlikely. The translational fusion however contains the full 5’UTR of malT, harboring the H-NS binding site. We thus hypothesized that AHL QS signaling represses H-NS-mediated activation of malT translation.

AHL enhances expression of cnu

The effect of AHL on the expression of H-NS was characterized in two different ways. First we evaluated the effect of AHL on the fluorescence of a GFP-H-NS fusion protein. Secondly we measured the effect of AHL on the relative abundance of the hns mRNA and that of its close paralogue stpA and the homologues hha and cnu by quantitative reverse transcription PCR (qRT-PCR). Fig. 3A shows the fluorescence of E. coli MG1655 carrying an hns-gfp fusion in place of chromosomal hns. The fluorescence of the GFP-H-NS fusion is unaffected by AHL. It must be

48

noted however, that the gfp-hns strain exhibits a partial hns phenotype. Specifically, the gfp-hns strain grows with a generation time intermediate of that of an isogenic MG1655 strain and an hns mutant, and the colonies appear mucoid similar to a hns mutant (data not shown). The gfp-hns strain is however able to ferment maltose, giving rise to red colonies on MacConkey maltose indicator agars, like the MG1655 parent strain, whereas the hns mutant is unable to ferment maltose and gives rise to white colonies, due to the inefficient translation of malT mRNA in the absence of H-NS (data not shown). Therefore, this fusion may not be subject to the same regulatory control as WT hns, and results obtained using this strain may not accurately reflect WT hns regulation. The relative abundance of hns and its paralogue stpA and the homologues hhA and cnu in the presence or absence of AHL were tested by qRT-PCR. Fig. 3B shows no effect of AHL on the relative mRNA levels of hns, stpA, or hha in WT E. coli MG1655. By contrast, AHL enhances the relative cnu abundance three fold (blue bars), and this effect is dependent on the presence of sdiA (green bars). By upregulating cnu, AHL could modulate the activity or specificity of H-NS.

contr

ol

AHL

Rela

tive f

luo

rescen

ce

arb

itra

ry u

nit

s

hns

stpA hh

Acnu

cnu

control

AHL

A B

WT sdiA

Fig. 3. The effect of AHL on expression of hns and homologues. (A) MG1655 hns-gfp was grown to an OD600 of 1 in the presence or absence of 5 µM oxo-C6-AHL. The relative fluorescence of GFP-H-NS was measured and normalized to MG1655 not carrying GFP, n=3. (B) MG1655 (blue bars) and an isogenic sdiA mutant (green bars) was grown to an OD600 of 0.5 in the presence or absence of 5 µM oxo-C6-AHL. The relative hns, stpA, hha and cnu mRNA levels were measured by qRT-PCR and normalized to the expression of the endogenous control cysG (Zhou et al., 2011). Error bars show standard deviations of at least three independent biological replicates. In order to mimic the AHL-induced upregulation of cnu observed in Fig. 3, we investigated the effect of overexpressing cnu from a low-copy-number plasmid on malT’-’lacZ β-galactosidase activity. Figure 4 shows that overexpression of cnu in a strain grown in the absence of AHL inhibits β-galactosidase activity to approximately the same extent as growth in the presence of AHL, suggesting that increased Cnu levels indeed negatively affect malT translation. However, overexpression of cnu combined with the presence of AHL further reduces the β-galactosidase activity, suggesting that AHL affects expression of malT even at elevated cellular levels of Cnu.

49

The importance of H-NS and its homologs for the AHL-mediated down regulation of the malT’-’lacZ fusion were assayed by constructing deletions of hns, stpA, hhA, cnu and a double deletion of both cnu and hha in the malT’-’lacZ reporter strain. Fig. 4 shows that deletion of hns completely abolishes β-galactosidase activity. This observation agrees with the previously described requirement of H-NS for efficient translation of the malT transcript (Park et al., 2010). Deleting the hns paralouge stpA does not affect the malT’-’lacZ fusion and no apparent phenotype is observed in this strain (Fig. 4). In the hha background, less β-galactosidase activity is observed, but the effect of AHL is similar to that of the wildtype strain. We note that the hha strain grows considerably slower compared to its parent strain, but the growth defect is not as pronounced as that of the hns strain (data not shown). These results may be explained by a previous observation that hha deficiency reduced H-NS expression and thereby caused a reduction in growth rate (Paytubi et al., 2004). Finally, although overexpression of cnu causes a drop in β-galactosidase activity similar to the drop caused by AHL, deletion of cnu does not abolish the inhibitory effect of AHL. Thus, AHL does not require cnu to repress malT, suggesting that AHL can also cause malT repression via another factor . The H-NS and Hha family proteins are known to be able to partially compensate for the loss of members of their respective families (Paytubi et al., 2004; Stoebel et al., 2008). Figure 4 shows that the double mutant deficient in both hha and cnu has significantly reduced β-galactosidase activity. This suggests that at least one of Cnu and Hha are required for efficient malT translation. It has previously been shown that a cnu hha double mutant expresses less of both stpA and hns (Paytubi et al., 2004). Whether the requirement for either cnu or hha for efficient malT translation is caused by the absence of these proteins to form heterodimers with H-NS or whether it is caused by the previously described effect of cnu and hha on H-NS abundance cannot be concluded. Figure 4 shows that overexpression of hns from a low-copy number plasmid in the wildtype reporter strain slightly reduces the inhibitory effect of AHL on malT’-’lacZ β-galactosidase activity, supporting the hypothesis that AHL repression of malT occurs at least partly through titration or downregulation of the H-NS pool that is available to bind malT mRNA.

50

Fig. 4. The effect of deletion and over expression of hns and the homologues stpA, hhA and cnu. MG1655 malT’-’lacZ (WT) and isogenic hns, stpA, hhA, cnu, cnu hhA mutant backgrounds or MG1655 malT’-’lacZ over expressing either hns or cnu from a pIDT plasmid were grown at 30°C in TB in with 5 µM oxo-C6-AHL or a solvent control. The relative β-galactosidase activity of the translational fusion was assayed at OD600 of 1 and normalized to a MG1655 lacZ strain. The β-galactosidase activity of the malT’-’lacZ fusion in WT control treated cells was set to one and the other values were normalized accordingly. Error bars show standard deviations of at least three independent biological replicates, except for the cnu hha double mutant, where n=2.

Discussion

In search for the regulatory pathway underlying the AHL-mediated repression of LamB expression in E. coli we discovered that malT, the positive regulator of lamB, is under negative translational control by AHL. This is in agreement with previous studies indicating that malT could be regulated by AHL and/or SdiA(Van Houdt et al., 2006; Wei et al., 2001). Although overexpression of the AHL receptor SdiA has been shown to regulate targets (Kanamaru et al., 2000; Sitnikov et al., 1996; Wang et al., 1991; Wei et al., 2001) that are not under regulation by chromosomally encoded sdiA in response to AHL (Dyszel et al., 2010), we here confirm that MalT, as well as the MalT-activated malE and lamB genes, are negatively regulated by oxo-C6-AHL and that this regulation is dependent on chromosomally encoded sdiA. We find that the AHL quorum sensing molecule oxo-C6-AHL upregulates the relative cnu mRNA levels, through the orphan AHL receptor SdiA, in E. coli MG1655. Cnu belongs to the Hha protein family of small proteins homologous to the N-terminal oligomerization domain of the global regulator H-NS, which can interact with H-NS and modulate its function (reviewed in (Dorman, 2014; Madrid et al., 2007)).

51

H-NS has a general affinity for all kinds of nucleic acids, but associates more strongly with dsDNA (Atlung and Ingmer, 1997). The affinity of H-NS for malT mRNA has not been measured, but the binding affinity of H-NS to two other RNAs, namely the small RNA DsrA and rpoS mRNA was found to be 3-4 fold lower than curved DNA (Brescia et al., 2004). Assuming that the majority of the highly abundant H-NS binds DNA, an increase in the approximately 130-fold less abundant Cnu (Paytubi et al., 2004; Wang et al., 2012) protein may exert a large effect on the fraction of H-NS not tightly bound to DNA. We propose that AHL upregulates Cnu to an extent where it interferes with H-NS association with malT mRNA and thus specifically inhibits the positive effect of H-NS on malT translation. However, growth in the presence of AHL represses malT translation even in a cnu mutant strain, suggesting that a redundant mediator of the AHL-effect has yet to be identified. Many of the genes that are under regulatory control by H-NS are induced by environmental signals, including osmolarity, temperature, and pH (reviewed in (Atlung and Ingmer, 1997)). Our results suggest that AHL quorum sensing, which is an environmental signal that enables E. coli to detect the presence of other Gram-negative species (Patankar and Gonzalez, 2009), affects at least a subset of H-NS targets through upregulation of the Hha family member cnu. Except from a previous observation that cnu expression was increased in an hha mutant background (Paytubi et al., 2004), we here identify the AHL-SdiA complex as the first positive regulator of cnu expression.

Acknowledgements

This work was supported in part by a research grant to S.L.S. from the Novo Nordisk Foundation. N.M.H.-K. was supported by the Danish National Research Foundation through the Center for Models of Life. We are grateful to Thomas J. Silhavy for bacterial strains and antibodies and we wish to thank him for inviting N.M.H.-K. to visit his lab. We also thank Anders Løbner-Olesen and Michael A. Sørensen for bacterial strains.

Materials and methods

Bacterial strains. The bacterial strains used in this study are Escherichia coli MG1655 and derivatives of this strain. They are listed in Table S1 in the supplemental material. NMHK35 was generated by P1 transduction of trpC Tn10 from ALO3938 into MG1655 and subsequently, Hns-gfp was P1 transduced from ALO1706 into this strain and selected for growth on M63 glucose plates. NMHK41 was generated by P1 transduction of the sdiA::cat allele from NMHK8, previously described (Hoyland-Kroghsbo et al., 2013), into the MG1655 recipient. The cat allele was removed using FLP-mediated recombination by expressing FLP from the plasmid pTL18 (Long et al., 2009). NMHK45 was generated by P1 transduction of trpC::kanR from the Keio collection (Baba et al., 2006) into MG1655, followed by testing for lacZ phenotype on MacConkey lactose agar, and P1 transduction of the WT proC allele from MC4100 and selection on M63 glucose plates, followed by testing for retention of the lacZ phenotype on MacConkey lactose agar. NMHK47 was made by P1 transduction of the lamB∆60′-′lacZ allele from BZR60 into NMHK45 and selection on M63 lactose plates (Emr and Silhavy, 1980). NMHK49 was generated by P1 transduction of the malT’-‘lacZ allele, encoding the first 9 amino acids of MalT fused to lacZ, from POP2481 into NMHK45 and selection on minimal lactose (Raibaud et al., 1991). NMHK50 was

52

made by P1 transduction of sdiA::cat into NMHK49 (Hoyland-Kroghsbo et al., 2013). NMHK76,77,78, and 79 were made by P1 transduction of the Keio alleles hns::kan, cnu::kan, hha::kan, and stpA::kan into NMHK49, respectively. NMHK97 was made by FLP-mediated recombination to remove kan from NMHK77, followed by P1 transduction of hha::kan. NMHK78 and 79 were made by electroporating pIDTcnu and pIDThns into NMHK49, respectively. NMHK69 was constructed by the method of Datsenko and Wanner, using the primers listed in Table S2 and MG1655 genomic DNA as template, followed by selection on M63 lactose for generating the transcriptional malT’-lacZ+ fusion in NMHK49 (Datsenko and Wanner, 2000). This construct harbors the iGEM SD BBa_J61107, in place of the native malT SD. All constructed strains were screened by PCR with primers flanking ORFs of the desired genes to identify mutants with gene replacements of the expected size. AHL preparation. N-(3-Oxohexanoyl)-L-homoserine lactone (K3007, Sigma) was dissolved in ethyl acetate (EA) acidified by 0.1% acetic acid and stored at -20°C. Glass culture tubes were coated with AHL by adding AHL to a final concentration of 5µM or the equivalent volume of EA alone and the EA was evaporated at RT until completely dry as described by Ghosh et al. (Ghosh et al., 2009). Bacterial growth conditions. For all experiments, E. coli MG1655 and its derivatives were grown from single colonies in TB medium (10 g tryptone and 5 g NaCl per liter) at 30°C, shaking at 220 rpm. Exponentially growing cultures were diluted approximately 107 into TB medium in glass tubes coated with AHL or control tubes treated with EA as described above and allowed to reach the desired cell density. Cell densities were measured by determining the optical density at 600 nm (OD600) on an Ultraspec 2100 Pro (Amersham Biosciences). One OD600 unit corresponds to a cell density of 1.1*109 CFU/ml. Ampicillin (100 µg/ml) was included as ne necessary to select for plasmid retention. The expression of pIDT plasmids were induced by 10 mM IPTG. Western blot. At OD600 0.75, 1 OD unit cells were harvested and boiled 5 min at 95°C in Tris sample buffer (67.5 mM Tris pH 6.8, 10% glycerol, 5% β-mercaptoethanol, 3% SDS, 0.05% bromophenol blue). The proteins were separated on a 10% SDS Bis Tris gel and blotted onto a nitrocellulose membrane (P/N 66485, Pall Life sciences). Equal loading was ensured by Ponceau S staining. The membrane was incubated O/N at 4°C with anti-LamB and anti-MalE antibody (Silhavy, TJ) diluted 1:25000 in TBST 5% skim milk. After three washes in TBST, the membrane was incubated 1h with anti-rabbit secondary antibody diluted 1:2000 in TBST 5% skim milk (A0545, Sigma). The membrane was developed using SuperSignal West Pico Chemiluminescent Substrate (34080, Thermo Scientific). β-galactosidase assays. At an OD600 of 1, one OD unit cells were harvested and resuspended in 1 ml Z-buffer supplemented with β-mercaptoethanol, chloroform and SDS, to permeabilize the cells. The relative β-galactosidase activity of strains carrying lacZ-fusions was assayed by o-nitrophenyl-β-D-galactopyranoside (ONPG) hydrolysis after 15 min by spectrophotometric readings in a FLUOstar Omega microplate reader (BMG Labtech). The relative β-galactosidase activities were normalized to that of MG1655 lacZ. Measurements of fluorescence. At an OD600 of 1, one OD unit cells were harvested and resuspended in 1 ml M63. 200 µl aliquots were immediately assayed for GFP activity by

53

spectrophotometric measurement of fluorescence in a FLUOstar Omega microplate reader (BMG Labtech). qRT-PCR. RNA was harvested from cells grown to OD600 0.5. RNA was extracted using TRI Reagent (T9424, Sigma) and the quality of the purified RNA was assessed using NanoDrop 1000 (Thermo Scientific). The RNA was DNAse treated using DNase I (EN0521, Thermo Scientific). cDNA was synthesized using RevertAid First Strand cDNA Synthesis Kit (K1622, Thermo Scientific). qPCR samples were prepared using SsoAdvanced Universal SYBR Green Supermix (172-5274, BIO-RAD). qPCR primers are listed in supplementary table S2. Delta CT values were calculated by 2−(CT ‘Gene of interest’ − CT ‘House keeping gene’).

54

Supplemental table S1

Bacterial strains and plasmids: Strains and plasmids Genotype Source or reference

MG1655 F-, rph-1 (Heine et al., 1988)

NMHK35 MG1655 gfp-hns This study

NMHK41 MG1655 ΔsdiA This study

NMHK45 MG1655 ΔlacZ This study

NMHK47 MG1655 ΔlacZ lamBΔ60’-‘lacZ This study

NMHK49 MG1655 ΔlacZ malT’-‘lacZ This study

NMHK50 MG1655 ΔlacZ malT’-‘lacZ sdiA::cat This study

NMHK69 MG1655 ΔlacZ malT’-lacZ+ This study

NMHK76 MG1655 ΔlacZ malT’-‘lacZ hns::kanR This study

NMHK77 MG1655 ΔlacZ malT’-‘lacZ cnu::kanR This study

NMHK78 MG1655 ΔlacZ malT’-‘lacZ hha::kanR This study

NMHK79 MG1655 ΔlacZ malT’-‘lacZ stpA::kanR This study

NMHK87 MG1655 ΔlacZ malT’-‘lacZ pIDTcnu This study

NMHK88 MG1655 ΔlacZ malT’-‘lacZ pIDThns This study

NMHK97 MG1655 ΔlacZ malT’-‘lacZ Δcnu hha::kanR This study

pKD46 Expressing the λ Red recombinase system (Datsenko and Wanner, 2000)

pTL18 Contains an IPTG-inducible FLP recombinase (Long et al., 2009)

pIDTcnu Contains the Ptac promotor, iGEM SD

BBa_J61107, the entire cnu ORF and lambda t1

transcriptional terminator, within an IDT smart

plasmid (IDT).

This study

pIDThns Contains the Ptac promotor, iGEM SD

BBa_J61107, the entire hns ORF and lambda t1

transcriptional terminator, within an IDT smart

plasmid (IDT).

This study

Supplemental table S2

Primers: Target gene Forward primer 5’-3’ Reverse primer 5’-3’

cysG qPCR TTGTCGGCGGTGGTGATGTC ATGCGGTGAACTGTGGAATAAACG

hns qPCR TGCTGCTGAAGTTGAAGAGC ACGTTTAGCTTTGGTGCCAG

stpA qPCR ACTAAAACCTGGACCGGTCA GATTTACCTTCTGCCAGCGC

hha qPCR GTTTACGTCGTTGCCAGACA CGGCTGAGTAAAATACCGCC

cnu qPCR ATGTATCGTGCTGCCGATCA GACATAGTGCCAGACGGACT

malT’-lacZ+ AAAAACGTCATCGCTTGCATTAGAAAGGTTTCT

GGCCGACCTTATAACCAAAAGAAGAGACTCAC

TAGATGACCATGATTACGGATTCAC

TTGAATGATGCAGAGATGTAAGCCGGATCTG

GCGCGTTATCCGGCTAAACTTAAGCGACTTC

ATTCACCTGAC

55

References

Atlung, T., and Ingmer, H. (1997). H-NS: a modulator of environmentally regulated gene expression.

Molecular microbiology 24, 7-17.

Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K.A., Tomita, M., Wanner,

B.L., and Mori, H. (2006). Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the

Keio collection. Molecular systems biology 2, 2006 0008.

Boos, W., and Shuman, H. (1998). Maltose/maltodextrin system of Escherichia coli: transport,

metabolism, and regulation. Microbiology and molecular biology reviews : MMBR 62, 204-229.

Brescia, C.C., Kaw, M.K., and Sledjeski, D.D. (2004). The DNA binding protein H-NS binds to and alters

the stability of RNA in vitro and in vivo. Journal of molecular biology 339, 505-514.

Datsenko, K.A., and Wanner, B.L. (2000). One-step inactivation of chromosomal genes in Escherichia

coli K-12 using PCR products. Proceedings of the National Academy of Sciences of the United States of

America 97, 6640-6645.

Dorman, C.J. (2014). H-NS-like nucleoid-associated proteins, mobile genetic elements and horizontal

gene transfer in bacteria. Plasmid 75C, 1-11.

Dyszel, J.L., Soares, J.A., Swearingen, M.C., Lindsay, A., Smith, J.N., and Ahmer, B.M. (2010). E. coli K-12

and EHEC genes regulated by SdiA. PloS one 5, e8946.

Emr, S.D., and Silhavy, T.J. (1980). Mutations affecting localization of an Escherichia coli outer

membrane protein, the bacteriophage lambda receptor. Journal of molecular biology 141, 63-90.

Ghosh, D., Roy, K., Williamson, K.E., Srinivasiah, S., Wommack, K.E., and Radosevich, M. (2009). Acyl-

homoserine lactones can induce virus production in lysogenic bacteria: an alternative paradigm for

prophage induction. Applied and environmental microbiology 75, 7142-7152.

Hoyland-Kroghsbo, N.M., Maerkedahl, R.B., and Svenningsen, S.L. (2013). A quorum-sensing-induced

bacteriophage defense mechanism. MBio 4.

Hughes, D.T., Terekhova, D.A., Liou, L., Hovde, C.J., Sahl, J.W., Patankar, A.V., Gonzalez, J.E., Edrington,

T.S., Rasko, D.A., and Sperandio, V. (2010). Chemical sensing in mammalian host-bacterial commensal

associations. Proceedings of the National Academy of Sciences of the United States of America 107, 9831-

9836.

Iino, T., and Mitani, M. (1967). Infection of Serratia marcescens by bacteriophage chi. Journal of

virology 1, 445-447.

Johansson, J., Dagberg, B., Richet, E., and Uhlin, B.E. (1998). H-NS and StpA proteins stimulate

expression of the maltose regulon in Escherichia coli. Journal of bacteriology 180, 6117-6125.

Kanamaru, K., Kanamaru, K., Tatsuno, I., Tobe, T., and Sasakawa, C. (2000). SdiA, an Escherichia coli

homologue of quorum-sensing regulators, controls the expression of virulence factors in

enterohaemorrhagic Escherichia coli O157:H7. Molecular microbiology 38, 805-816.

Lehnen, D., Blumer, C., Polen, T., Wackwitz, B., Wendisch, V.F., and Unden, G. (2002). LrhA as a new

transcriptional key regulator of flagella, motility and chemotaxis genes in Escherichia coli. Molecular

microbiology 45, 521-532.

Long, T., Tu, K.C., Wang, Y., Mehta, P., Ong, N.P., Bassler, B.L., and Wingreen, N.S. (2009). Quantifying

the integration of quorum-sensing signals with single-cell resolution. PLoS biology 7, e68.

Madrid, C., Balsalobre, C., Garcia, J., and Juarez, A. (2007). The novel Hha/YmoA family of nucleoid-

associated proteins: use of structural mimicry to modulate the activity of the H-NS family of proteins.

Molecular microbiology 63, 7-14.

Meynell, E.W. (1961). A Phage, Phichi, Which Attacks Motile Bacteria. J Gen Microbiol 25, 253-&.

Michael, B., Smith, J.N., Swift, S., Heffron, F., and Ahmer, B.M. (2001). SdiA of Salmonella enterica is a

LuxR homolog that detects mixed microbial communities. Journal of bacteriology 183, 5733-5742.

56

Ng, W.L., and Bassler, B.L. (2009). Bacterial quorum-sensing network architectures. Annual review of

genetics 43, 197-222.

Park, H.S., Ostberg, Y., Johansson, J., Wagner, E.G., and Uhlin, B.E. (2010). Novel role for a bacterial

nucleoid protein in translation of mRNAs with suboptimal ribosome-binding sites. Genes & development

24, 1345-1350.

Patankar, A.V., and Gonzalez, J.E. (2009). Orphan LuxR regulators of quorum sensing. FEMS

microbiology reviews 33, 739-756.

Paytubi, S., Madrid, C., Forns, N., Nieto, J.M., Balsalobre, C., Uhlin, B.E., and Juarez, A. (2004). YdgT, the

Hha paralogue in Escherichia coli, forms heteromeric complexes with H-NS and StpA. Molecular

microbiology 54, 251-263.

Raibaud, O., Vidal-Ingigliardi, D., and Kolb, A. (1991). Genetic studies on the promoter of malT, the gene

that encodes the activator of the Escherichia coli maltose regulon. Research in microbiology 142, 937-942.

Randall-Hazelbauer, L., and Schwartz, M. (1973). Isolation of the bacteriophage lambda receptor from

Escherichia coli. Journal of bacteriology 116, 1436-1446.

Sertic, V., and Boulgakov, N.A. (1936). Specific bacteriophages for flagellated bacterial varieties. Cr Soc

Biol 123, 887-888.

Sharma, V.K., Bearson, S.M., and Bearson, B.L. (2010). Evaluation of the effects of sdiA, a luxR

homologue, on adherence and motility of Escherichia coli O157 : H7. Microbiology 156, 1303-1312.

Sitnikov, D.M., Schineller, J.B., and Baldwin, T.O. (1996). Control of cell division in Escherichia coli:

regulation of transcription of ftsQA involves both rpoS and SdiA-mediated autoinduction. Proceedings of

the National Academy of Sciences of the United States of America 93, 336-341.

Stoebel, D.M., Free, A., and Dorman, C.J. (2008). Anti-silencing: overcoming H-NS-mediated repression

of transcription in Gram-negative enteric bacteria. Microbiology 154, 2533-2545.

Van Houdt, R., Aertsen, A., Moons, P., Vanoirbeek, K., and Michiels, C.W. (2006). N-acyl-L-homoserine

lactone signal interception by Escherichia coli. FEMS microbiology letters 256, 83-89.

Wang, M., Weiss, M., Simonovic, M., Haertinger, G., Schrimpf, S.P., Hengartner, M.O., and von Mering,

C. (2012). PaxDb, a database of protein abundance averages across all three domains of life. Molecular &

cellular proteomics : MCP 11, 492-500.

Wang, X.D., de Boer, P.A., and Rothfield, L.I. (1991). A factor that positively regulates cell division by

activating transcription of the major cluster of essential cell division genes of Escherichia coli. The EMBO

journal 10, 3363-3372.

Wei, Y., Lee, J.M., Smulski, D.R., and LaRossa, R.A. (2001). Global impact of sdiA amplification revealed

by comprehensive gene expression profiling of Escherichia coli. Journal of bacteriology 183, 2265-2272.

Zhou, K., Zhou, L., Lim, Q., Zou, R., Stephanopoulos, G., and Too, H.P. (2011). Novel reference genes for

quantifying transcriptional responses of Escherichia coli to protein overexpression by quantitative PCR.

BMC molecular biology 12, 18.

57

5. Manuscript B: Interspecies quorum sensing accelerates E. coli entry into stationary phase

58

Interspecies quorum sensing accelerates E. coli entry into stationary phase Nina Molin Høyland-Kroghsbo, Linda Hove Christensen and Sine Lo Svenningsen

Institute of Biology, University of Copenhagen, Copenhagen, Denmark

Abstract

Bacteria have to adapt quickly to changes in their environment. In E. coli, multiple stimulatory inputs including starvation, changes in temperature, osmotic stress and antibiotics activate the general stress response, which is essential to survive these challenges. A hallmark of stationary phase is the induction of rpoS, which encodes the stationary phase alternative sigma factor responsible for orchestrating the general stress response. Contrary to previous studies, we find that the AHL quorum sensing molecule 3-oxo-C6-HSL induces δS in E. coli. Using an rpoS-mCherry fusion, we show that AHL induces δS earlier in the growth phase. δS stimulates formation of persister cells, which are a transiently multidrug tolerant subpopulation of an otherwise antibiotic-sensitive isogenic bacterial population. Importantly, we find that AHL signaling increases the fraction persister cells and that this is dependent on the AHL receptor sdiA as well as rpoS.

Introduction

In a process called quorum sensing (QS), bacteria communicate, monitor population density, and control important collective behaviors, including bioluminescence, biofilm formation, and virulence. QS depends on the production, release, and detection of diffusible signaling molecules called autoinducers (AI). When an extracellular AI threshold concentration is achieved, the bacterial population responds by inducing large numbers of genes in synchrony, thereby coordinating behavior across the population, enabling the bacteria to act in unison (Ng and Bassler, 2009). E. coli produces and responds to the QS molecule AI-2, which is a universal interspecies AI that is widespread among the bacterial kingdom (Pereira et al., 2013). N-acyl-L-homoserine lactones (AHLs) are another major class of intraspecies AI’s, which are produced by Gram-negative species. AHLs are composed of a homoserine lactone ring with an acyl side chain, which varies in length from C4 to C12 and can be modified at C3 (Ng and Bassler, 2009). Usually, one species produces and detects only a single or a few types of AHL. Exceptions include E. coli and other enterobacteria, which are able to detect a wide range of AHLs through the AHL receptor SdiA, but they are themselves unable to synthesize AHLs as they lack a cognate synthase (Smith et al., 2011; Soares and Ahmer, 2011; Yao et al., 2006). Why these species possess the orphan AHL receptor SdiA is debated, but it would enable them to detect and act upon the presence of other Gram-negative species, which gives cues about the environment. SdiA is required for the pathogenic E. coli O157:H7 to colonize cattle (Hughes et al., 2010). This requirement for SdiA was proposed to be caused by the positive regulation by SdiA of the glutamate-dependent acid resistance system (Dyszel et al., 2010) required to survive the gastric acidic environment.

59

We previously reported that E. coli regulates a bacteriophage defense mechanism in response to AHL quorum sensing, thus increasing its chances of surviving a virulent bacteriophage attack (Hoyland-Kroghsbo et al., 2013). Inspired by our discovery that E. coli utilizes AHL QS in its risk assessment of phage infection, we hypothesize that AHL may be perceived as an environmental stress signal and that this in turn may activate more general stress mechanisms in E. coli. We thus hypothesized that that SdiA may regulate the general stress response in E. coli. In E. coli δS is the master regulator of the general stress response, enabling the bacterium to cope with a range of environmental stresses. δ

S associates with RNA polymerase and promotes transcription of the rpoS regulon, which includes up to 10% of the E. coli genome (Landini et al., 2014). Stress signals including changes in temperature, pH, osmolarity, and nutrient availability trigger δS production or activity, which is often accompanied by a reduction in growth rate. The general stress response renders cells resistant or tolerant to multiple stresses, and not only to the specific stress that triggered the response (Hengge-Aronis, 2002). rpoS is an important positive regulator of persister cell formation in E. coli (Tkachenko et al., 2014). Persister cells are a transiently multidrug tolerant subpopulation of an otherwise antibiotic-sensitive isogenic bacterial population (reviewed in (Maisonneuve and Gerdes, 2014)). The phenotypic trait of persistence was observed 70 years ago by J. W. Bigger, who suggested that dormant non-replicating cells are the basis of this phenomenon (Bigger, 1944). (p)ppGpp is the central regulatory molecule in the stringent response, a stress response induced by amino acid starvation. Amino acid starvation results in accumulation of uncharged tRNAs in the cell, which in turn activates (p)ppGpp synthesis by RelA. SpoT is responsible for (p)ppGpp synthesis in response to other stress signals (reviewed in (Magnusson et al., 2005; Starosta et al., 2014)). Korch et al. found that the high frequency of persistent cells in a hipA7 mutant is dependent on functional relA and spoT genes and they proposed that the stringent response may form the basis of the persistent stage (Korch et al., 2003). The model for (p)ppGpp-mediated formation of persisters is based on stochastic induction of (p)ppGpp (Maisonneuve et al., 2013), which inhibits exopolyphosphatase, the cellular enzyme that degrades polyphosphate. The resulting increase in polyphosphate activates the Lon protease to degrade all type II antitoxins. The freed toxins then inhibit translation and cell growth and thereby result in persisters (reviewed in (Maisonneuve and Gerdes, 2014)). The signaling pathways triggering the persister phenotype are highly complex and involves integration of multiple stimulatory inputs, including the stringent- and SOS DNA-damage responses. Environmental signals such as nutrient limitation, subinhibitory concentrations of antibiotics, oxidative stress, heat shock and DNA damaging agents all stimulate persister cell formation (reviewed in (Helaine and Kugelberg, 2014; Maisonneuve and Gerdes, 2014)). The stringent response and the general stress response are intertwined in that (p)ppGpp is involved in regulating both the production and activity of δS (Magnusson et al., 2005). Due to the lack of (p)ppGpp, a relA spoT strain expresses very little δS and consequently has multiple amino acid requirements and is hyper sensitive to stress including heat and cold shock, osmotic stress, and antibiotics (Hengge-Aronis, 2002; Maisonneuve et al., 2013; Xiao et al., 1991). Conversely, δS

60

appears to positively affect the formation of persister cells, as the persister fraction of a population of rpoS mutant cells is diminished (Tkachenko 2014), and the persister fraction increases along with δ

S activity as a population enters stationary phase (figure S2D, (Maisonneuve et al., 2013)). A single study claims the opposite effect, namely that a culture of rpoS mutant cells is almost completely resistant to ampicillin (Hong et al. 2012). QS has been found to increase persister cell formation in P. aeruginosa and Streptococcus mutans (Leung and Levesque, 2012; Moker et al., 2010). The involvement of SdiA in E. coli resistance to antibiotics is controversial: overexpression of SdiA from a plasmid was found to increase antibiotic tolerance (Rahmati et al., 2002) whereas the presence of AHL QS molecules did not increase antibiotic tolerance in wildtype cells harboring chromosomally encoded SdiA (Dyszel et al., 2010). We noted that the AHL receptor SdiA and δ

S regulate a common set of stress related genes; both regulators activate the glutamate-dependent acid resistance (gad) system (Dong and Schellhorn, 2010; Dyszel et al., 2010; Hughes et al., 2010; Van Houdt et al., 2006; Waterman and Small, 2003) (Patten et al., 2004), repress genes involved in formation of flagella (Dyszel et al., 2010; Sharma et al., 2010) (Dong and Schellhorn, 2009; Patten et al., 2004; Uchiyama et al., 2010), and inhibit expression of LEE encoded virulence factors in the enterohemorrhagic group of pathogenic E. coli EHEC (Dong and Schellhorn, 2010; Hughes et al., 2010). Additionally, both SdiA and δS are important for the EHEC O157:H7 to survive passage through the bovine gastrointestinal tract (Hughes et al., 2010; Price et al., 2000). Therefore, we sought to investigate whether AHL quorum sensing through the receptor SdiA affects the master regulator of the general stress response, δS. We found that AHL QS signaling leads to an accumulation of the master regulator of the general stress response, δS, in late log phase, suggesting that AHL accelerates entry into stationary phase. Importantly, we found that AHL enhances the persister cell frequency in E. coli. This finding could have potential clinical implications, as regrowth of persisters after antibiotic treatment may lead to reoccurrence of infections (reviewed in (Helaine and Kugelberg, 2014)). Inhibition of AHL quorum sensing may reduce E. coli persister cell formation and thereby reduce relapse of infections after antibiotic treatment.

Results

AHL quorum sensing signaling leads to accumulation of δS

We investigated the effect of the AHL 3-oxo-C6-HSL on the relative δS levels in E. coli K-12 MG1655. As rpoS gene expression does not always correlate with δ

S protein levels, due to extensive post-transcriptional regulation of rpoS translation and δS stability (Battesti et al., 2011; Hengge-Aronis, 2002), we chose to test the effect of AHL on δS protein levels by western blotting, using a δS-specific antibody. Our previous results showed AHL-mediated regulation most clearly when a cell culture started from a single colony has shown exponential balanced growth at 30 °C until an OD600 ranging from 0.5 to 1.0 (Hoyland-Kroghsbo et al., 2013). We thus evaluated the relative δS expression under these

61

conditions. Figure 1 shows that δS protein levels are upregulated in response to AHL, and that this regulation is dependent on the orphan AHL receptor SdiA.

Fig. 1 AHL signaling results in δS protein accumulation Western blot of total cell lysate from WT and sdiA mutant cells. The cells were grown to an OD600 of 0.75 in the presence or absence of 5 µM 3-oxo-C6-HSL. The data is representative of four independent experiments.

AHL signaling induces δS accumulation earlier in growth phase

In order to investigate the timing of AHL-mediated induction of δS, we tested the relative fluorescence generated from expression of a chromosomal rpoS-mCherry fusion (Maisonneuve et al., 2013), in cultures grown in the presence or absence of AHL. Figure 2 shows that AHL signaling indeed affects the timing of δS accumulation. When grown in the presence of AHL, E. coli accumulates δS-mCherry fusion protein earlier in the growth phase compared to cells treated with the solvent control, suggesting that AHL signaling induces δS production, and thereby accelerates E. coli entry into stationary phase.

62

Rela

tive f

luo

res

ce

nce

arb

itra

ry u

nit

s

Fig. 2 AHL signaling modulates timing and extent of δS accumulation Relative fluorescence of MG1655 rpoS-mCherry in response to AHL. Diluted cultures of MG1655 rpoS-mCherry were grown at 30 °C in TB in the presence or absence of 5 µM 3-oxo-C6-HSL for 12 h until OD600 of 0.25. Aliquots of these cultures were transferred to a 96 well mictrotiterplate and the relative fluorescence was monitored every 15 min during incubation at 30 °C with intermittent shaking. The graph shows averages of four biological replicates and is representative of three independent experiments. Error bars indicate SD.

AHL increases persister cell formation in E. coli

Because δS has been shown to positively affect persister cell formation in E. coli (Tkachenko et al., 2014), and we find that AHL leads to an accumulation of δS, we hypothesized that AHL may affect persister cell formation in E. coli. Figure 3 shows the effect of the AHL quorum sensing signal 3-oxo-C6-HSL on the fraction of WT and isogenic sdiA and rpoS mutants of E. coli that are able to resuscitate after 1h and 3h treatment with the β-lactam ampicillin, respectively. We observed a large day-to-day variation in the fraction of persisters that are able to resume growth after ampicillin treatment. This observation is consistent with observations made by E. Maisonneuve (personal communication). In order to compare data obtained from different days, we chose to show the fold change in persister fraction of each AHL-treated culture compared to a culture treated with the solvent control. Figure 3 shows that AHL enhances the fraction of persisters that can form a colony following both 1h and 3h ampicillin exposure, The effect depends on the presence of both sdiA and rpoS. Of note, we find that the effect of AHL on the fraction of persister cells is more pronounced at OD600 0.5 compared to OD600 1.0 (data not shown).

63

Fig. 3 Effect of AHL on persister cell formation in E. coli. WT, sdiA and rpoS mutant E. coli cells were grown in the presence of 5 µM 3-oxo-C6-HSL or a solvent control. At OD600 0.5 exponentially growing cells were exposed to 100 mg/ml ampicillin. After 1 or 3 hours the cells were washed, serially diluted, and plated in the absence of antibiotics to enumerate the CFU. Each data point represents the CFU of one independent culture treated with AHL, divided by the CFU of the control treated culture arising from the same single colony. The horizontal lines show the mean. For WT and sdiA n=12 and for rpoS n=6.

Discussion Quorum sensing enables bacteria to estimate population density and in some cases species composition. The orphan AHL QS receptor SdiA enables E. coli and other enteric bacteria to detect a wide range of AHLs, allowing them to sense the presence of the majority of Gram-negative AHL-producing species (Smith et al., 2011). In high-cell-density mixed species environments, competition between species likely induces various survival mechanisms. Detection of AHLs may be one mechanism whereby E. coli could prepare itself for competition and the stress imposed by rapid exhaustion of nutrients and accumulation of toxic metabolites, in a dense mixed-species environment. Therefore, we hypothesized that AHL could induce a general stress response in E. coli. Indeed, we find that E. coli responds to the AHL quorum sensing signaling molecule 3-oxo-C6-HSL by upregulating δS, the master regulator of the general stress response in E. coli (Battesti et al., 2011). This regulation is dependent on the AHL receptor SdiA. AHL signaling induces δS

accumulation earlier in the growth phase, suggesting that AHL accelerates E. coli entry into stationary phase. As δS positively affects persister cell formation (Tkachenko et al., 2014), AHL-induced onset of δS accumulation, may explain our observation of an increased persister cell

64

fraction in AHL-treated WT E. coli. The dependence of rpoS for the stimulatory effect of AHL on persister cell formation further supports this interpretation. The involvement of SdiA in rpoS regulation has been investigated previously with contradictory results: Huisman and Kolter found that addition of homoserine lactone caused increased δ

S protein levels measured by western blot. The direct effect of δS protein levels was not tested in an sdiA mutant, but an assay for δS-dependent KPII catalase activity showed that an sdiA mutant expressed similar levels of catalase activity compared to WT cells, which goes in line with our finding that the δ

S level is not affected in the sdiA mutant, in the absence of AHL (Huisman and Kolter, 1994). However, Sitnikov et al. were unable to reproduce the finding that homoserine lactone increases δS protein levels using transcriptional and translational rpoS-lacZ fusions (Sitnikov et al., 1996). The fusions used by Sitnikov et al. originally constructed by (Loewen et al., 1993) did however not include the most upstream of the rpoS promoters, which is located inside the nlpD gene, and drives the majority of rpoS transcription. Furhtermore, the the nlpD promoters, which drive nlpD-rpoS operon transcription accounting for low-level of rpoS expression during exponential growth were also not included (Lange et al., 1995; Takayanagi et al., 1994). Thus, the key regulatory sites may have been missed. Our finding that AHL QS signaling increases the persister fraction in E. coli is in line with previous reports of QS-induction of persister cell formation in the Gram-negative opportunistic human pathogen Pseudomonas aeruginosa (P. aeruginosa) and the Gram-positive Streptococcus mutans, which is involved in tooth decay, suggesting that this phenomenon may be widespread (Leung and Levesque, 2012; Moker et al., 2010). In addition to evaluating the effect of QS on P. aeruginosa persister formation, Moker et al. tested the effect in E. coli. They added spent medium from the 3-oxo-C12-HSL-producing P. aeruginosa stain PA14 to E. coli in early logarithmic growth for 1.5 h and subsequently evaluated the persister fraction after treatment with carbenicillin and piperacillin. They found no effect of the spent medium in E. coli persister formation (Moker et al., 2010). Both ampicillin, carbenicillin and piperacillin belong to the extended-spectrum beta-lactam antibiotics and they all work by inhibiting cell wall synthesis, so the use of different antibiotics is unlikely to explain the contradictory results. The reason why Monker et al. do not find an AHL-mediated effect on persister cell formation in E. coli could potentially be due to the different type of AHL they tested or likely a consequence of the relatively short exposure to spent medium. Our experience shows that balanced growth for many generations in the presence of AHL is important for consistent AHL-mediated regulation. Plasmid based overexpression of sdiA has been found to upregulate the AcrAB efflux pump and thereby increase E. coli tolerance to multiple antibiotics (Rahmati et al., 2002), as measured by the E-Test for evaluating the minimum inhibitory concentration of antimicrobial agents on solid agar (Baker et al., 1991). The E-test thus measures the general antibiotic tolerance rather than the presence of rare persister cells. Dyszel et al. later confirmed these results using plasmid-encoded sdiA. Importantly, they further demonstrated that sdiA expressed from its natural chromosomal location does not regulate acrAB expression in response to AHL, nor does it affect the susceptibility

65

of E. coli to antibiotics, measured by E-Test and in motility agar (Dyszel et al., 2010). Thus, we do not expect that the observed AHL-effect is related to regulation of the AcrAB efflux pump. In Salmonella enterica serovar Typhimurium, sdiA expression is differentially regulated, with two bursts of expression: one in exponential phase and one in early stationary phase (Turnbull et al., 2012). The authors additionally found that rpoS is responsible for stationary phase expression of sdiA. If rpoS is also responsible for regulating sdiA in E. coli, there is a positive regulatory feedback loop between rpoS and sdiA, which may give rise to an accumulation of both proteins in late logarithmic phase or early stationary phase, which could be enhanced further in the presence of AHL. It is tempting to speculate that such a regulatory feedback loop may underlie our observations that AHL QS leads to enhanced persister levels in E. coli and it may further explain the overlapping phenotypes imposed by the general stress response and AHL signaling, including enhanced acid resistance, inhibition of flagella formation and repression of LEE genes (Dong and Schellhorn, 2009, 2010; Dyszel et al., 2010; Hughes et al., 2010; Patten et al., 2004; Sharma et al., 2010; Uchiyama et al., 2010; Van Houdt et al., 2006; Waterman and Small, 2003), as well as the requirement of rpoS and sdiA for survival through the bovine GI tract (Hughes et al., 2010; Price et al., 2000). Interestingly, Goo et al. has shown that three species of Burkholderia induce functions of importance for survival in stationary phase in response to endogenously produced AHL (Goo et al. 2012). In line with this finding, we propose that a major role of the orphan SdiA-receptor in E. coli is to enable E. coli cells to anticipate stresses related to growth in a high-cell-density mixed-species environment by accelerating δS production in response to the accumulation of quorum-sensing signals produced by other gram-negative bacteria. Hereby we propose an answer for the long standing question as to why E. coli possesses a receptor for AHL QS molecules that it does not itself produce, namely that it enables E. coli to use interspecies quorum sensing signals as a stress-anticipation signal.

Acknowledgements

This work was supported in part by a research grant to S.L.S. from the Novo Nordisk Foundation. N.M.H.-K. was supported by the Danish National Research Foundation through the Center for Models of Life. We are grateful to Kenn Gerdes and Anders Løbner-Olesen for bacterial strains.

Materials and Methods

Bacterial strains. The bacterial strains used in this study are Escherichia coli MG1655 and derivatives of this strain. They are listed in Table S1 in the supplemental material. NMHK41 was generated by P1 transduction of the sdiA::cat allele from NMHK8, previously described (Hoyland-Kroghsbo et al., 2013), into the MG1655 recipient. cat was removed using FLP-mediated recombination by expressing FLP from the plasmid pTL18 (Long et al., 2009). NMHK118 MG1655 rpoS::Tn10 was generated by P1 transduction of rpoS359::Tn10 from RH90 (Lange and

66

Hengge-Aronis, 1991) into MG1655. LHC2 MG1655 rpoS-mCherry was constructed by P1 transduction of the rpoS-mCherry allele previously described (Maisonneuve et al., 2013) into NMHK118 and selection for loss of tetracycline resistance by growth in the presence of fusaric acid (Bochner et al., 1980). All constructed strains were screened by PCR with primers flanking ORFs of the desired genes to identify mutants with gene replacements of the expected size. AHL preparation. N-(3-Oxohexanoyl)-L-homoserine lactone (K3007, Sigma) was dissolved in ethyl acetate (EA) acidified by 0.1% acetic acid and stored at -20°C. Glass culture tubes were coated with AHL by adding AHL to a final concentration of 5µM or the equivalent volume of EA alone and the EA was evaporated at RT until completely dry as described by Ghosh et al. (Ghosh et al., 2009). Bacterial growth conditions. For all experiments, E. coli MG1655 and its derivatives were grown from single colonies in TB medium (10 g tryptone and 5 g NaCl per liter) at 30°C, shaking at 220 rpm. Exponentially growing cultures were diluted approximately 107 into TB medium in glass tubes coated with AHL or control tubes treated with EA as described above and allowed to reach the desired cell density. Cell densities were measured by determining the optical density at 600 nm (OD600) on an Ultraspec 2100 Pro (Amersham Biosciences). One OD600 unit corresponds to a cell density of 1.1*109 CFU/ml. Western blot. At OD600 0.75, 1 OD unit cells were harvested and boiled 5 min at 95°C in Tris sample buffer (67.5 mM Tris pH 6.8, 10% glycerol, 5% β-mercaptoethanol, 3% SDS, 0.05% bromophenol blue). The proteins were separated on a 10% SDS Bis Tris gel and blotted onto a nitrocellulose membrane (P/N 66485, Pall Life sciences). Equal loading was ensured by Ponceau S staining. The membrane was incubated O/N at 4°C with anti-δS antibody (WP009, Neoclone) diluted 1:2000 in TBST 5% skim milk. After three washes in TBST, the membrane was incubated 1h with anti-mouse secondary antibody diluted 1:2000 in TBST 5% skim milk (Sigma A2554). The membrane was developed using SuperSignal West Pico Chemiluminescent Substrate (34080, Thermo Scientific). Persister assay. At OD600 0.5 exponentially growing cells were exposed to 100 mg/ml ampicillin. After 1 or 3 hours the cells were washed and serially diluted in M63, and plated in the absence of antibiotics to enumerate the CFU. Measurement of fluorescence. Diluted cultures of MG1655 rpoS-mCherry were grown at 30 °C in TB in the presence or absence of 5 µM 3-oxo-C6-HSL for 12 h until OD600 of 0.25. For each experiment four AHL-treated and four control cultures were grown and aliquots of each culture were transferred in four technical replicates to a 96 well mictrotiterplate and the relative fluorescence was monitored every 15 min during incubation at 30 °C with intermittent shaking in a FLUOstar Omega microplate reader (BMG Labtech).

67

Supplemental table S1

Bacterial strains and plasmids are listed in table S1: Strains and plasmid Genotype Source or reference

MG1655 F-, rph-1 (Heine et al., 1988)

NMHK41 MG1655 ∆sdiA This study

NMHK118 MG1655 rpoS Tn10 This study

LHC2 NMHK118 rpoS-mCherry This study

pTL18 Contains an IPTG-inducible FLP recombinase (Long et al., 2009)

68

References

Baker, C.N., Stocker, S.A., Culver, D.H., and Thornsberry, C. (1991). Comparison of the E Test to agar

dilution, broth microdilution, and agar diffusion susceptibility testing techniques by using a special

challenge set of bacteria. Journal of clinical microbiology 29, 533-538.

Battesti, A., Majdalani, N., and Gottesman, S. (2011). The RpoS-mediated general stress response in

Escherichia coli. Annual review of microbiology 65, 189-213.

Bigger, J.W. (1944). Treatment of staphylococcal infections with penicillin - By intermittent sterilisation.

Lancet 2, 497-500.

Bochner, B.R., Huang, H.C., Schieven, G.L., and Ames, B.N. (1980). Positive selection for loss of

tetracycline resistance. Journal of bacteriology 143, 926-933.

Dong, T., and Schellhorn, H.E. (2009). Control of RpoS in global gene expression of Escherichia coli in

minimal media. Molecular genetics and genomics : MGG 281, 19-33.

Dong, T., and Schellhorn, H.E. (2010). Role of RpoS in virulence of pathogens. Infection and immunity

78, 887-897.

Dyszel, J.L., Soares, J.A., Swearingen, M.C., Lindsay, A., Smith, J.N., and Ahmer, B.M. (2010). E. coli K-12

and EHEC genes regulated by SdiA. PloS one 5, e8946.

Ghosh, D., Roy, K., Williamson, K.E., Srinivasiah, S., Wommack, K.E., and Radosevich, M. (2009). Acyl-

homoserine lactones can induce virus production in lysogenic bacteria: an alternative paradigm for

prophage induction. Applied and environmental microbiology 75, 7142-7152.

Heine, H.G., Francis, G., Lee, K.S., and Ferenci, T. (1988). Genetic analysis of sequences in maltoporin

that contribute to binding domains and pore structure. Journal of bacteriology 170, 1730-1738.

Helaine, S., and Kugelberg, E. (2014). Bacterial persisters: formation, eradication, and experimental

systems. Trends in microbiology 22, 417-424.

Hengge-Aronis, R. (2002). Signal transduction and regulatory mechanisms involved in control of the

sigma(S) (RpoS) subunit of RNA polymerase. Microbiology and molecular biology reviews : MMBR 66, 373-

395, table of contents.

Hoyland-Kroghsbo, N.M., Maerkedahl, R.B., and Svenningsen, S.L. (2013). A quorum-sensing-induced

bacteriophage defense mechanism. MBio 4.

Hughes, D.T., Terekhova, D.A., Liou, L., Hovde, C.J., Sahl, J.W., Patankar, A.V., Gonzalez, J.E., Edrington,

T.S., Rasko, D.A., and Sperandio, V. (2010). Chemical sensing in mammalian host-bacterial commensal

associations. Proceedings of the National Academy of Sciences of the United States of America 107, 9831-

9836.

Huisman, G.W., and Kolter, R. (1994). Sensing starvation: a homoserine lactone--dependent signaling

pathway in Escherichia coli. Science 265, 537-539.

Korch, S.B., Henderson, T.A., and Hill, T.M. (2003). Characterization of the hipA7 allele of Escherichia

coli and evidence that high persistence is governed by (p)ppGpp synthesis. Molecular microbiology 50,

1199-1213.

Landini, P., Egli, T., Wolf, J., and Lacour, S. (2014). sigmaS, a major player in the response to

environmental stresses in Escherichia coli: role, regulation and mechanisms of promoter recognition.

Environmental microbiology reports 6, 1-13.

Lange, R., Fischer, D., and Hengge-Aronis, R. (1995). Identification of transcriptional start sites and the

role of ppGpp in the expression of rpoS, the structural gene for the sigma S subunit of RNA polymerase in

Escherichia coli. Journal of bacteriology 177, 4676-4680.

69

Lange, R., and Hengge-Aronis, R. (1991). Identification of a central regulator of stationary-phase gene

expression in Escherichia coli. Molecular microbiology 5, 49-59.

Leung, V., and Levesque, C.M. (2012). A stress-inducible quorum-sensing peptide mediates the

formation of persister cells with noninherited multidrug tolerance. Journal of bacteriology 194, 2265-2274.

Loewen, P.C., von Ossowski, I., Switala, J., and Mulvey, M.R. (1993). KatF (sigma S) synthesis in

Escherichia coli is subject to posttranscriptional regulation. Journal of bacteriology 175, 2150-2153.

Long, T., Tu, K.C., Wang, Y., Mehta, P., Ong, N.P., Bassler, B.L., and Wingreen, N.S. (2009). Quantifying

the integration of quorum-sensing signals with single-cell resolution. PLoS biology 7, e68.

Magnusson, L.U., Farewell, A., and Nystrom, T. (2005). ppGpp: a global regulator in Escherichia coli.

Trends in microbiology 13, 236-242.

Maisonneuve, E., Castro-Camargo, M., and Gerdes, K. (2013). (p)ppGpp controls bacterial persistence

by stochastic induction of toxin-antitoxin activity. Cell 154, 1140-1150.

Maisonneuve, E., and Gerdes, K. (2014). Molecular mechanisms underlying bacterial persisters. Cell

157, 539-548.

Moker, N., Dean, C.R., and Tao, J. (2010). Pseudomonas aeruginosa increases formation of multidrug-

tolerant persister cells in response to quorum-sensing signaling molecules. Journal of bacteriology 192,

1946-1955.

Ng, W.L., and Bassler, B.L. (2009). Bacterial quorum-sensing network architectures. Annual review of

genetics 43, 197-222.

Patten, C.L., Kirchhof, M.G., Schertzberg, M.R., Morton, R.A., and Schellhorn, H.E. (2004). Microarray

analysis of RpoS-mediated gene expression in Escherichia coli K-12. Molecular genetics and genomics :

MGG 272, 580-591.

Pereira, C.S., Thompson, J.A., and Xavier, K.B. (2013). AI-2-mediated signalling in bacteria. FEMS

microbiology reviews 37, 156-181.

Price, S.B., Cheng, C.M., Kaspar, C.W., Wright, J.C., DeGraves, F.J., Penfound, T.A., Castanie-Cornet,

M.P., and Foster, J.W. (2000). Role of rpoS in acid resistance and fecal shedding of Escherichia coli O157:H7.

Applied and environmental microbiology 66, 632-637.

Rahmati, S., Yang, S., Davidson, A.L., and Zechiedrich, E.L. (2002). Control of the AcrAB multidrug efflux

pump by quorum-sensing regulator SdiA. Molecular microbiology 43, 677-685.

Sharma, V.K., Bearson, S.M., and Bearson, B.L. (2010). Evaluation of the effects of sdiA, a luxR

homologue, on adherence and motility of Escherichia coli O157 : H7. Microbiology 156, 1303-1312.

Sitnikov, D.M., Schineller, J.B., and Baldwin, T.O. (1996). Control of cell division in Escherichia coli:

regulation of transcription of ftsQA involves both rpoS and SdiA-mediated autoinduction. Proceedings of

the National Academy of Sciences of the United States of America 93, 336-341.

Smith, J.L., Fratamico, P.M., and Yan, X. (2011). Eavesdropping by bacteria: the role of SdiA in

Escherichia coli and Salmonella enterica serovar Typhimurium quorum sensing. Foodborne pathogens and

disease 8, 169-178.

Soares, J.A., and Ahmer, B.M. (2011). Detection of acyl-homoserine lactones by Escherichia and

Salmonella. Current opinion in microbiology 14, 188-193.

Starosta, A.L., Lassak, J., Jung, K., and Wilson, D.N. (2014). The bacterial translation stress response.

FEMS microbiology reviews.

Takayanagi, Y., Tanaka, K., and Takahashi, H. (1994). Structure of the 5' upstream region and the

regulation of the rpoS gene of Escherichia coli. Molecular & general genetics : MGG 243, 525-531.

Tkachenko, A.G., Kashevarova, N.M., Karavaeva, E.A., and Shumkov, M.S. (2014). Putrescine controls

the formation of Escherichia coli persister cells tolerant to aminoglycoside netilmicin. FEMS microbiology

letters.

Turnbull, A.L., Kim, W., and Surette, M.G. (2012). Transcriptional regulation of sdiA by cAMP-receptor

protein, LeuO, and environmental signals in Salmonella enterica serovar Typhimurium. Canadian journal of

microbiology 58, 10-22.

70

Uchiyama, J., Nobue, Y., Zhao, H., Matsuzaki, H., Nagahama, H., Matsuoka, S., Matsumoto, K., and Hara,

H. (2010). Involvement of sigmaS accumulation in repression of the flhDC operon in acidic phospholipid-

deficient mutants of Escherichia coli. Microbiology 156, 1650-1660.

Van Houdt, R., Aertsen, A., Moons, P., Vanoirbeek, K., and Michiels, C.W. (2006). N-acyl-L-homoserine

lactone signal interception by Escherichia coli. FEMS microbiology letters 256, 83-89.

Waterman, S.R., and Small, P.L. (2003). Transcriptional expression of Escherichia coli glutamate-

dependent acid resistance genes gadA and gadBC in an hns rpoS mutant. Journal of bacteriology 185, 4644-

4647.

Xiao, H., Kalman, M., Ikehara, K., Zemel, S., Glaser, G., and Cashel, M. (1991). Residual guanosine 3',5'-

bispyrophosphate synthetic activity of relA null mutants can be eliminated by spoT null mutations. The

Journal of biological chemistry 266, 5980-5990.

Yao, Y., Martinez-Yamout, M.A., Dickerson, T.J., Brogan, A.P., Wright, P.E., and Dyson, H.J. (2006).

Structure of the Escherichia coli quorum sensing protein SdiA: activation of the folding switch by acyl

homoserine lactones. Journal of molecular biology 355, 262-273.

71

6. Discussion

This PhD thesis provides evidence that AHL quorum sensing signals emitted from other Gram-

negative species are an important regulator of transient tolerance in E. coli. Quorum sensing-

regulation of transient tolerance would allow E. coli to quickly and reversibly respond to high-cell-

densities of AHL-emitting Gram-negative species in their environment, thus protecting it from a

diverse range of environmental stresses. Specifically, we find that AHL signaling activates a

bacteriophage defense mechanism, thus protecting E. coli from attacks by specific phages.

Additionally we find that AHL signaling induces accumulation of δS earlier in the growth phase and

hence accelerates the entry into stationary phase, a physiological state where E. coli is tolerant to a

broad range of environmental stresses. Finally, AHL upregulates the expression of Cnu, an H-NS

binding protein that could mediate a change in the target specificity of this global histone-like

regulatory protein, providing an additional way to modify the bacteriums gene expression in

response to AHL.

AHL as a stress-anticipation signal AHL perception by E. coli allows the bacterium to detect the presence and abundance of other

Gram-negative species in the environment. A crowded mixed species environment may quickly be

exhausted of nutrients and oxygen and consequently, toxic waste products would accumulate, all of

which imposes stress upon the bacteria. AHL detection by E. coli may be perceived as a warning of

exhausting resources and could allow E. coli to prepare itself for starvation and stress in due time.

Indeed, we find that AHL QS mediates stationary-phase preparation by inducing δS, the stationary

phase and stress-induced sigma factor, which orchestrates the general stress response (reviewed in

(Battesti et al., 2011; Landini et al., 2014)).

In support of our findings, the Gram-negative Burkholderia species pseudomallei, glumae and

thailandensis were found to emit the AHL C8-HSL and in response to its accumulation, they

produced oxalate, which protected stationary phase cells from ammonia-induced base toxicity.

Upon entry into stationary phase, mutants deficient in C8-HSL production were rapidly killed (Goo

et al., 2012). Thus, AHL signaling allows these Burkholderia species to anticipate stationary-phase

72

stress and respond in a preventive manner by producing a metabolite that neutralizes stationary

phase toxic waste products. Additionally, the rice pathogen Burkholderia glumae was further found

to restrict glucose uptake and nucleotide metabolism in response to C8-HSL. The authors proposed

that QS-mediated metabolic repression provides survival benefits under nutrient limitations as a

consequence of overcrowding (An et al., 2014).

QS-mediated induction of mechanisms required to cope with stationary phase stress may very well

be a general phenomenon among quorum sensing species.

Common targets for SdiA and δS

Some previously described targets of AHL-mediated regulation and the general stress response,

governed by δS, overlap: both regulators activate the glutamate-dependent acid resistance system

(Dong and Schellhorn, 2010; Dyszel et al., 2010; Hughes et al., 2010; Patten et al., 2004; Van

Houdt et al., 2006; Waterman and Small, 2003), repress genes involved in formation of flagella

synthesis (Dong and Schellhorn, 2009; Dyszel et al., 2010; Patten et al., 2004; Sharma et al., 2010;

Uchiyama et al., 2010), and inhibit expression of LEE-encoded virulence factors in EHEC (Dong

and Schellhorn, 2010; Hughes et al., 2010). Additionally, both SdiA and δS are important for

EHEC O157:H7 to survive passage through the bovine gastrointestinal tract (Hughes et al., 2010;

Price et al., 2000). We speculate that the AHL-mediated upregulation of δS may underlie or

contribute to the previously described AHL-mediated regulation of these targets.

Quorum sensing-mediated regulation of phage defense Since our discovery that AHL QS signaling increases E. coli transient tolerance to attack by phages

λvir and χ, by down regulation of their respective receptors, another case of QS-mediated phage

defense via. down regulation of phage receptors has been discovered. In collaboration with our

laboratory, Tan et al. (in preparation), recently uncovered a case of an AHL QS-mediated

downregulation of the phage receptor OmpK in the marine fish-pathogen Vibrio anguillarum,

pointing towards QS-mediated regulation of phage receptors as a general bacterial phage defense

mechanism.

73

Cnu-mediated regulation Apart from cnu expression being increased in an hha mutant background (Paytubi et al., 2004), to

the best of our knowledge, we here describe the first positive regulator of cnu. We find that AHL

QS signaling via the AHL-receptor SdiA leads to a three-fold increase in cnu expression and we

propose that this contributes to the AHL-mediated posttranscriptional regulation of malT, which is

supported by our finding that overexpression of cnu mimics the regulatory effect of AHL on malT.

Accumulation of Cnu may sequester H-NS from its role as an activator of malT translation. Since

AHL-mediated regulation of malT translation occurs even in the absence of cnu or both cnu and

hha, it appears that AHL-mediated regulation of malT can occur via two or more redundant

pathways.

Interestingly, both sdiA and hha deletions were found to increase expression of the fibronectin-

binding curli subunit csgA as well as EHEC O157:H7 adherence to human HEp-2 cells and a double

sdiA and hha mutant had a synergistic effect on csgA expression and cell adhesion (Sharma et al.,

2010). This points towards a cooperative regulatory mechanism of SdiA and Hha on csgA

inhibition. We have found by qRT-PCR analysis that AHL downregulates csgA expression (our

unpublished observation). δS has been found to relieve H-NS-mediated transcriptional silencing of

csgA (Olsen et al., 1993). It is attractive to speculate whether the SdiA- and AHL-mediated

downregulation of csgA expression could be due to the AHL-mediated upregulation of δS, or if it

could be caused by AHL-mediated regulation of Cnu, which in turn modulates H-NS transcriptional

silencing of csgA, or both.

Only a few cases of Cnu-mediated regulation has been described to date. Cnu negatively modulates

expression of S. typhimurium pathogenicity island 2, ensuring appropriate timing of its expression

during infection (Coombes et al., 2005). In E. coli, Cnu and Hha were found to form complexes

with H-NS on the origin of replication (oriC) and this interaction was proposed to be important for

replication (Kim et al., 2005). Another mechanism affected by upregulation of Cnu in E. coli is

Rho-dependent termination. First it was found that overexpression of Cnu restores polarity in rho

and nusG mutants, which are otherwise impaired in Rho-dependent transcription termination. It was

also found that mutations in cnu and hha greatly adds to the growth defect of a rho mutant (Saxena

and Gowrishankar, 2011). In a later report, the effect of Cnu on Rho-dependent termination of the

tryptophanase tnaLAB operon was investigated. It was found that deletion of cnu and hha led to

derepression the tna operon downstream of the Rho terminator. Interestingly, the θ subunit of DNA

74

polymerase III, holE was also found to contribute to the Rho-dependent termination of the tnaLAB

operon. Structure comparisons of HolE, Cnu, and Hha showed striking similarity. The authors

proposed that HolE serves as a molecular backup for Cnu (Dietrich et al., 2014). Potentially, the

AHL-mediated upregulation of cnu expression may affect the regulation of these three described

Cnu-mediated functions.

Further experiments

Immediate questions arise regarding the further experiments, which may give answers to the

unsolved issues in Manuscripts A and B.

In order to determine if the AHL-mediated repression of malT is due to an effect that is solely

dependent on the ability of H-NS to interact with malT mRNA, we will test a mutant of the malT-

lacZ translational fusion containing four basepair substitutions in the H-NS binding site, known to

abolish H-NS binding (Park et al., 2010). As very little LacZ activity is predicted from this mutant,

it will be tested in a mlc mutant background, in which malT transcription is greatly enhanced

(Decker et al., 1998). We have observed a high upregulation of LacZ activity of the malT-lacZ

fusion in a mlc background, and the negative effect of AHL is unaltered in this mutant (data not

shown). If the AHL-effect is undetectable in this strain background, it will support our hypothesis

that the SdiA-AHL complex mediates its effect on malT-lacZ via modulation of H-NS activity. The

malT-lacZ fusion is repressed two-fold in the cnu hha double mutant (Fig. 4, Manuscript A). We

will test whether overexpression of H-NS in the cnu hha background restores malT-lacZ expression

to that of the WT. If so, it supports our hypothesis that cnu affects malT-lacZ transcription by

altering the target specificity of H-NS. Finally, as δS has been proposed to be responsible for

overcoming H-NS mediated transcriptional silencing in bacteria undergoing stress (reviewed in

(Stoebel et al., 2008)), AHL-mediated upregulation of δS may additionally affect the ability of H-

NS to enhance malT translation. Therefore, we will also determine whether rpoS is required for the

AHL-mediated repression of malT-lacZ transcription, by testing the fusion in an rpoS mutant

background.

With regards to the study of AHL-accelerated entry into stationary phase, we will first test the effect

of AHL on the expression of the rpoS-Mcherry construct in a sdiA background to determine if SdiA

is required for the AHL-mediated regulation, as is predicted from our western blot of δS in sdiA

mutant cells (Figure 1, manuscript B).

75

Interestingly, we observe a window in the growth phase where the stimulatory effect of AHL on

rpoS-Mcherry expression is present. This phenomenon may be due to a growth phase dependent

expression of SdiA (Turnbull et al., 2012), the dependence of other factors that are expressed during

this window in the growth phase, or it may be related to induction of the AI-2 QS system, which in

our experience is induced around the onset of the AHL-mediated effect on δS accumulation (data

not shown). Testing the AHL-response on rpoS-mCherry expression in a background deficient in

the AI-2 synthase luxS, would allow us to determine whether the AHL-effect is AI-2-dependent. If

the AHL-induction of δS is dependent on the universal QS signal AI-2, it would allow E. coli to

integrate information about not only the presence of AHL-producing Gram negative species, but

also the presence of its own species, as well as other the Gram positive- and Gram negative AI-2

emitting bacteria, into its regulation of the general stress response. Moreover, in order to determine

if the AHL-mediated increase in persister cells is specific to ampicillin, or if AHL confers transient

antibiotic tolerance towards antibiotics with other modes of action, we will test whether AHL

increases the fraction of persisters that are able to regrow after treatment with the DNA gyrase

inhibitor ciprofloxacin. If we find that AHL increases the persister frequency in this assay, it

supports our hypothesis that AHL increases E. coli’s transient tolerance towards a broad range of

environmental stresses.

Another important question is to what extend AHL increases the fitness of E. coli in stationary

phase, through its positive effect on the stationary phase sigma factor δS. We will mix equal

numbers of WT and an sdiA::cat mutant resistant to chloramphenicol and test if AHL affects their

competitive fitness in late stationary phase by enumerating CFU on plates with and without

chloramphenicol and deducing the composition of the viable cells present in late stationary phase. If

the WT strain outgrows the sdiA::cat mutant in the presence of AHL, this would further support our

hypothesis that AHL protects E. coli form stationary phase stress.

Another unsolved question concerns the nature of the regulatory pathway underlying AHL-

mediated δS accumulation. Since the stress alarmone (p)ppGpp positively affects δS expression as

well as persister cell formation, we will test the effect of AHL on persister cell formation and δS

expression in relA spoT and relA spoT sdiA mutant backgrounds, which we have constructed. If we

find that RelA and SpoT are required for the AHL-mediated effect on persister formation and δS

expression, we will test if AHL regulates the expression of relA and/or spoT by qRT-PCR.

76

Another potential regulatory pathway for AHL-regulation of δS expression is via an AHL-regulated

Cnu-modulatory effect on H-NS-mediated destabilization on rpoS mRNA. By qRT-PCR

measurements on the relative rpoS abundance at different time points after rifampicin treatment, we

can test if AHL affects rpoS mRNA stability.

Ultimately, in order to characterize the full regulatory pathways underlying the AHL-mediated

mechanisms we have identified in manuscript A and B, we will test if SdiA binds directly to the

promoters of putative direct targets. We are currently setting up an SdiA EMSA in order to test if

SdiA associates directly with cnu and rpoS promoters, and if we find that AHL activates expression

of relA and/or spoT by qRT-PCR, the promoters of these genes will also be tested. This approach

would allow us to potentially expand the currently modest number of described direct SdiA-targets

in the literature.

In light of two reports of a positive effect of Cnu on Rho-dependent transcriptional termination

(Dietrich et al., 2014; Saxena and Gowrishankar, 2011), it is highly appealing to speculate whether,

AHL-mediated upregulation of cnu affects Rho-dependent transcription termination. We have at

hand a reporter of Rho efficiency, consisting of the trp operon, followed by a Rho-dependent

terminator and the lac operon fused directly downstream (Guarente and Beckwith, 1978). In this

construct, lack of Rho-dependent termination at the end of trp allows read-through of the lac

operon. Thus increased Rho activity correlates with reduced lac transcription. Testing this reporter

in a genetic background with a rho mutant of reduced activity, will allow us to evaluate to what

extent AHL-mediated upregulation of cnu is able to improve termination efficiency in the rho

mutant. This approach may uncover a completely new regulatory pathway whereby AHL can

regulate gene expression.

Major unanswered questions within the field of SdiA research remain. Only a few direct binding

sites for SdiA have been identified, and no consensus binding motif has yet been recognized.

Therefore, global mapping of SdiA binding sites in the E. coli chromosome would be highly

desired. This could be achieved by ChIP-Seq, a method that combines chromatin

immunoprecipitation (ChIP) with DNA sequencing (Seq).

In light of our discovery that the orphan AHL-receptor SdiA allows E. coli to anticipate stress in

response to AHL signaling, it raises the question as to what extend this phenomenon occurs in the

77

species Salmonella, Klebsiella, and Shigella, which harbor an orphan SdiA receptor and also to

what extend AHL-regulates stress anticipation across the AHL-producing Gram-negative species.

7. Conclusion This PhD thesis provides a key answer to why E. coli listens in on QS, from other Gram-negative

species, through the orphan AHL receptor SdiA. Namely, detection of interspecies AHL quorum

sensing by E. coli serves to anticipate- and adapt to environmental stress.

Specifically, the first quorum sensing-regulated bacteriophage defense mechanism was uncovered

(Article 1). Investigating the underlying regulatory mechanism led to the finding that AHL-

mediated QS upregulated the expression of cnu, encoding an Hha-family protein that interacts with

H-NS and potentially modulates its translational enhancement on malT mRNA. SdiA was identified

as the first inducer of cnu expression (Manuscript A), apart from a previous observation that it is

upregulated in an hha mutant (Paytubi et al., 2004).

Inspired by the discovery that E. coli utilized AHL signals as a risk-indicator of phage infection and

activated phage defenses accordingly, it was hypothesized that AHL QS signaling may be perceived

as a more universal environmental stress signal, allowing E. coli to activate the general stress

response. This was indeed demonstrated by the finding that AHL signaling induced accumulation of

the stationary phase alternative sigma factor δS, which orchestrates the general stress response

(reviewed in (Battesti et al., 2011; Landini et al., 2014)). The induction of δS led to increased

persister cell formation in the AHL treated population, and this was dependent on sdiA and rpoS,

suggesting that AHL may regulate persister cell formation via its positive effect on δS (Manuscript

B).

78

8. Perspectives

Uncovering how bacteria perceive environmental signals and how they interpret these, in order to

constantly adapt to changes in their environment, is important for understanding not only microbial

ecology but also bacterial pathogenesis and it provides cues as to how we might interfere with these

systems, in order to prevent undesirable bacterial behavior. The production and detection of AIs

enable bacteria to estimate the abundance and species complexity of a microbial community and

forms the basis of key bacterial group-behaviors such as virulence.

Inhibitors of bacterial cell-cell communication are emerging as promising alternatives to

conventional antibiotics in the battle against multi-drug resistant bacteria. Disrupting bacterial cell-

cell communication may disarm pathogens and thus render them harmless, allowing patient’s

immune systems to clear infections. This approach, is expected to minimally disturb the natural

microbiome and, as it does not kill bacteria, it would additionally put less selective pressure on the

population, which would reduce the risk of resistant subpopulations emerging (reviewed in (Marx,

2014; Rutherford and Bassler, 2012)).

Considering our discovery, that AHL protects E. coli from phage infection, it is attractive to

speculate that a combination of QS-inhibitory drugs with medically administered phages, so called

phage therapy, may render pathogenic E. coli harmless and additionally make it more susceptible to

phage attack.

In light of our finding that AHL signaling leads to accumulation of δS, the master regulator of the

general stress response, and additionally leads to induced persister cell formation, QS-inhibitory

drugs may furthermore generally inhibit the tolerance and fitness of pathogens as well as reduce the

frequency of persister cell formation. This may especially be relevant in the case of chronic or

recurrent infections, where patients receive repeated treatments with antibiotics of increasing

potency, and where the risk of selection for true antibiotic resistant strains is high. Ultimately,

treatment with QS-inhibitory drugs in combination with classical antibiotics may reduce the

occurrence of relapse of infections that are due to persister cells. This may in turn reduce the

number of antibiotic treatments a given patient would need.

79

9. References An, J.H., Goo, E., Kim, H., Seo, Y.S., and Hwang, I. (2014). Bacterial quorum sensing and metabolic slowing in

a cooperative population. Proceedings of the National Academy of Sciences of the United States of

America.

Atlung, T., and Ingmer, H. (1997). H-NS: a modulator of environmentally regulated gene expression.

Molecular microbiology 24, 7-17.

Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K.A., Tomita, M., Wanner, B.L.,

and Mori, H. (2006). Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio

collection. Molecular systems biology 2, 2006 0008.

Baker, C.N., Stocker, S.A., Culver, D.H., and Thornsberry, C. (1991). Comparison of the E Test to agar

dilution, broth microdilution, and agar diffusion susceptibility testing techniques by using a special

challenge set of bacteria. Journal of clinical microbiology 29, 533-538.

Bassler, B.L., and Losick, R. (2006). Bacterially speaking. Cell 125, 237-246.

Battesti, A., Majdalani, N., and Gottesman, S. (2011). The RpoS-mediated general stress response in

Escherichia coli. Annual review of microbiology 65, 189-213.

Bigger, J.W. (1944). Treatment of staphylococcal infections with penicillin - By intermittent sterilisation.

Lancet 2, 497-500.

Bochner, B.R., Huang, H.C., Schieven, G.L., and Ames, B.N. (1980). Positive selection for loss of tetracycline

resistance. Journal of bacteriology 143, 926-933.

Boos, W., and Shuman, H. (1998). Maltose/maltodextrin system of Escherichia coli: transport, metabolism,

and regulation. Microbiology and molecular biology reviews : MMBR 62, 204-229.

Brescia, C.C., Kaw, M.K., and Sledjeski, D.D. (2004). The DNA binding protein H-NS binds to and alters the

stability of RNA in vitro and in vivo. Journal of molecular biology 339, 505-514.

Brussow, H., and Hendrix, R.W. (2002). Phage genomics: Small is beautiful. Cell 108, 13-16.

Chaudhuri, R.R., and Henderson, I.R. (2012). The evolution of the Escherichia coli phylogeny. Infection,

genetics and evolution : journal of molecular epidemiology and evolutionary genetics in infectious diseases

12, 214-226.

Coombes, B.K., Wickham, M.E., Lowden, M.J., Brown, N.F., and Finlay, B.B. (2005). Negative regulation of

Salmonella pathogenicity island 2 is required for contextual control of virulence during typhoid.

Proceedings of the National Academy of Sciences of the United States of America 102, 17460-17465.

Court, D.L., Oppenheim, A.B., and Adhya, S.L. (2007). A new look at bacteriophage lambda genetic

networks. Journal of bacteriology 189, 298-304.

Datsenko, K.A., and Wanner, B.L. (2000). One-step inactivation of chromosomal genes in Escherichia coli K-

12 using PCR products. Proceedings of the National Academy of Sciences of the United States of America

97, 6640-6645.

Decker, K., Plumbridge, J., and Boos, W. (1998). Negative transcriptional regulation of a positive regulator:

the expression of malT, encoding the transcriptional activator of the maltose regulon of Escherichia coli, is

negatively controlled by Mlc. Molecular microbiology 27, 381-390.

80

Dietrich, M., Pedro, L., Garcia, J., Pons, M., Huttener, M., Paytubi, S., Madrid, C., and Juarez, A. (2014).

Evidence for moonlighting functions of the theta subunit of Escherichia coli DNA polymerase III. Journal of

bacteriology 196, 1102-1112.

Dong, T., and Schellhorn, H.E. (2009). Control of RpoS in global gene expression of Escherichia coli in

minimal media. Molecular genetics and genomics : MGG 281, 19-33.

Dong, T., and Schellhorn, H.E. (2010). Role of RpoS in virulence of pathogens. Infection and immunity 78,

887-897.

Dorman, C.J. (2014). H-NS-like nucleoid-associated proteins, mobile genetic elements and horizontal gene

transfer in bacteria. Plasmid 75C, 1-11.

Dyszel, J.L., Soares, J.A., Swearingen, M.C., Lindsay, A., Smith, J.N., and Ahmer, B.M. (2010). E. coli K-12 and

EHEC genes regulated by SdiA. PloS one 5, e8946.

Emr, S.D., and Silhavy, T.J. (1980). Mutations affecting localization of an Escherichia coli outer membrane

protein, the bacteriophage lambda receptor. Journal of molecular biology 141, 63-90.

Ghosh, D., Roy, K., Williamson, K.E., Srinivasiah, S., Wommack, K.E., and Radosevich, M. (2009). Acyl-

homoserine lactones can induce virus production in lysogenic bacteria: an alternative paradigm for

prophage induction. Applied and environmental microbiology 75, 7142-7152.

Gomez, P., and Buckling, A. (2011). Bacteria-phage antagonistic coevolution in soil. Science 332, 106-109.

Goo, E., Majerczyk, C.D., An, J.H., Chandler, J.R., Seo, Y.S., Ham, H., Lim, J.Y., Kim, H., Lee, B., Jang, M.S., et

al. (2012). Bacterial quorum sensing, cooperativity, and anticipation of stationary-phase stress. Proceedings

of the National Academy of Sciences of the United States of America 109, 19775-19780.

Gottesman, M.E., and Weisberg, R.A. (2004). Little lambda, who made thee? Microbiology and molecular

biology reviews : MMBR 68, 796-813.

Guarente, L.P., and Beckwith, J. (1978). Mutant RNA polymerase of Escherichia coli terminates transcription

in strains making defective rho factor. Proceedings of the National Academy of Sciences of the United

States of America 75, 294-297.

Hall, A.R., Scanlan, P.D., Morgan, A.D., and Buckling, A. (2011). Host-parasite coevolutionary arms races give

way to fluctuating selection. Ecology letters 14, 635-642.

Hargreaves, K.R., Kropinski, A.M., and Clokie, M.R. (2014). What does the talking?: quorum sensing

signalling genes discovered in a bacteriophage genome. PloS one 9, e85131.

Heine, H.G., Francis, G., Lee, K.S., and Ferenci, T. (1988). Genetic analysis of sequences in maltoporin that

contribute to binding domains and pore structure. Journal of bacteriology 170, 1730-1738.

Helaine, S., and Kugelberg, E. (2014). Bacterial persisters: formation, eradication, and experimental

systems. Trends in microbiology 22, 417-424.

Hengge-Aronis, R. (2002). Signal transduction and regulatory mechanisms involved in control of the

sigma(S) (RpoS) subunit of RNA polymerase. Microbiology and molecular biology reviews : MMBR 66, 373-

395, table of contents.

Hirakawa, H., and Tomita, H. (2013). Interference of bacterial cell-to-cell communication: a new concept of

antimicrobial chemotherapy breaks antibiotic resistance. Frontiers in microbiology 4, 114.

Hong, S.H., Wang, X., O'Connor, H.F., Benedik, M.J., and Wood, T.K. (2012). Bacterial persistence increases

as environmental fitness decreases. Microbial biotechnology 5, 509-522.

Hoyland-Kroghsbo, N.M., Maerkedahl, R.B., and Svenningsen, S.L. (2013). A quorum-sensing-induced

bacteriophage defense mechanism. MBio 4.

Hughes, D.T., Terekhova, D.A., Liou, L., Hovde, C.J., Sahl, J.W., Patankar, A.V., Gonzalez, J.E., Edrington, T.S.,

Rasko, D.A., and Sperandio, V. (2010). Chemical sensing in mammalian host-bacterial commensal

associations. Proceedings of the National Academy of Sciences of the United States of America 107, 9831-

9836.

Huisman, G.W., and Kolter, R. (1994). Sensing starvation: a homoserine lactone--dependent signaling

pathway in Escherichia coli. Science 265, 537-539.

Iino, T., and Mitani, M. (1967). Infection of Serratia marcescens by bacteriophage chi. Journal of virology 1,

445-447.

81

Johansson, J., Dagberg, B., Richet, E., and Uhlin, B.E. (1998). H-NS and StpA proteins stimulate expression of

the maltose regulon in Escherichia coli. Journal of bacteriology 180, 6117-6125.

Kanamaru, K., Kanamaru, K., Tatsuno, I., Tobe, T., and Sasakawa, C. (2000). SdiA, an Escherichia coli

homologue of quorum-sensing regulators, controls the expression of virulence factors in

enterohaemorrhagic Escherichia coli O157:H7. Molecular microbiology 38, 805-816.

Kaper, J.B., Nataro, J.P., and Mobley, H.L. (2004). Pathogenic Escherichia coli. Nature reviews Microbiology

2, 123-140.

Kim, M.S., Bae, S.H., Yun, S.H., Lee, H.J., Ji, S.C., Lee, J.H., Srivastava, P., Lee, S.H., Chae, H., Lee, Y., et al.

(2005). Cnu, a novel oriC-binding protein of Escherichia coli. Journal of bacteriology 187, 6998-7008.

Korch, S.B., Henderson, T.A., and Hill, T.M. (2003). Characterization of the hipA7 allele of Escherichia coli

and evidence that high persistence is governed by (p)ppGpp synthesis. Molecular microbiology 50, 1199-

1213.

Labrie, S.J., Samson, J.E., and Moineau, S. (2010). Bacteriophage resistance mechanisms. Nature reviews

Microbiology 8, 317-327.

Landini, P., Egli, T., Wolf, J., and Lacour, S. (2014). sigmaS, a major player in the response to environmental

stresses in Escherichia coli: role, regulation and mechanisms of promoter recognition. Environmental

microbiology reports 6, 1-13.

Lange, R., Fischer, D., and Hengge-Aronis, R. (1995). Identification of transcriptional start sites and the role

of ppGpp in the expression of rpoS, the structural gene for the sigma S subunit of RNA polymerase in

Escherichia coli. Journal of bacteriology 177, 4676-4680.

Lange, R., and Hengge-Aronis, R. (1991). Identification of a central regulator of stationary-phase gene

expression in Escherichia coli. Molecular microbiology 5, 49-59.

Lederberg, E.M. (1950). Lysogenicity in Escherichia coli strain K-12. Microbial Genetics Bulletin 1, 5-8.

Lee, J.H., and Lee, J. (2010). Indole as an intercellular signal in microbial communities. FEMS microbiology

reviews 34, 426-444.

Lehnen, D., Blumer, C., Polen, T., Wackwitz, B., Wendisch, V.F., and Unden, G. (2002). LrhA as a new

transcriptional key regulator of flagella, motility and chemotaxis genes in Escherichia coli. Molecular

microbiology 45, 521-532.

Leimbach, A., Hacker, J., and Dobrindt, U. (2013). E. coli as an All-Rounder: The Thin Line Between

Commensalism and Pathogenicity. Curr Top Microbiol 358, 3-32.

Leung, V., and Levesque, C.M. (2012). A stress-inducible quorum-sensing peptide mediates the formation of

persister cells with noninherited multidrug tolerance. Journal of bacteriology 194, 2265-2274.

Lieb, M. (1953a). The establishment of lysogenicity in Escherichia coli. Journal of bacteriology 65, 642-651.

Lieb, M. (1953b). Studies on lysogenization in Escherichia coli. Cold Spring Harbor symposia on quantitative

biology 18, 71-73.

Little, J.W., and Mount, D.W. (1982). The SOS regulatory system of Escherichia coli. Cell 29, 11-22.

Loewen, P.C., von Ossowski, I., Switala, J., and Mulvey, M.R. (1993). KatF (sigma S) synthesis in Escherichia

coli is subject to posttranscriptional regulation. Journal of bacteriology 175, 2150-2153.

Long, T., Tu, K.C., Wang, Y., Mehta, P., Ong, N.P., Bassler, B.L., and Wingreen, N.S. (2009). Quantifying the

integration of quorum-sensing signals with single-cell resolution. PLoS biology 7, e68.

Madrid, C., Balsalobre, C., Garcia, J., and Juarez, A. (2007). The novel Hha/YmoA family of nucleoid-

associated proteins: use of structural mimicry to modulate the activity of the H-NS family of proteins.

Molecular microbiology 63, 7-14.

Magnusson, L.U., Farewell, A., and Nystrom, T. (2005). ppGpp: a global regulator in Escherichia coli. Trends

in microbiology 13, 236-242.

Maisonneuve, E., Castro-Camargo, M., and Gerdes, K. (2013). (p)ppGpp controls bacterial persistence by

stochastic induction of toxin-antitoxin activity. Cell 154, 1140-1150.

Maisonneuve, E., and Gerdes, K. (2014). Molecular mechanisms underlying bacterial persisters. Cell 157,

539-548.

82

Maisonneuve, E., Shakespeare, L.J., Jorgensen, M.G., and Gerdes, K. (2011). Bacterial persistence by RNA

endonucleases. Proceedings of the National Academy of Sciences of the United States of America 108,

13206-13211.

Marx, V. (2014). Cell communication: stop the microbial chatter. Nature 511, 493-497.

Melechen, N.E., and Go, G. (1980). Induction of lambdoid prophages by amino acid deprivation: differential

inducibility; role of recA. Molecular & general genetics : MGG 180, 147-155.

Meynell, E.W. (1961). A Phage, Phichi, Which Attacks Motile Bacteria. J Gen Microbiol 25, 253-&.

Michael, B., Smith, J.N., Swift, S., Heffron, F., and Ahmer, B.M. (2001). SdiA of Salmonella enterica is a LuxR

homolog that detects mixed microbial communities. Journal of bacteriology 183, 5733-5742.

Moker, N., Dean, C.R., and Tao, J. (2010). Pseudomonas aeruginosa increases formation of multidrug-

tolerant persister cells in response to quorum-sensing signaling molecules. Journal of bacteriology 192,

1946-1955.

Nealson, K.H., Platt, T., and Hastings, J.W. (1970). Cellular control of the synthesis and activity of the

bacterial luminescent system. Journal of bacteriology 104, 313-322.

Ng, W.L., and Bassler, B.L. (2009). Bacterial quorum-sensing network architectures. Annual review of

genetics 43, 197-222.

Olsen, A., Arnqvist, A., Hammar, M., Sukupolvi, S., and Normark, S. (1993). The RpoS sigma factor relieves

H-NS-mediated transcriptional repression of csgA, the subunit gene of fibronectin-binding curli in

Escherichia coli. Molecular microbiology 7, 523-536.

Oppenheim, A.B., Kobiler, O., Stavans, J., Court, D.L., and Adhya, S. (2005). Switches in bacteriophage

lambda development. Annual review of genetics 39, 409-429.

Park, H.S., Ostberg, Y., Johansson, J., Wagner, E.G., and Uhlin, B.E. (2010). Novel role for a bacterial

nucleoid protein in translation of mRNAs with suboptimal ribosome-binding sites. Genes & development

24, 1345-1350.

Patankar, A.V., and Gonzalez, J.E. (2009). Orphan LuxR regulators of quorum sensing. FEMS microbiology

reviews 33, 739-756.

Patten, C.L., Kirchhof, M.G., Schertzberg, M.R., Morton, R.A., and Schellhorn, H.E. (2004). Microarray

analysis of RpoS-mediated gene expression in Escherichia coli K-12. Molecular genetics and genomics :

MGG 272, 580-591.

Paytubi, S., Madrid, C., Forns, N., Nieto, J.M., Balsalobre, C., Uhlin, B.E., and Juarez, A. (2004). YdgT, the Hha

paralogue in Escherichia coli, forms heteromeric complexes with H-NS and StpA. Molecular microbiology

54, 251-263.

Pennisi, E. (2013). The CRISPR craze. Science 341, 833-836.

Pereira, C.S., Thompson, J.A., and Xavier, K.B. (2013). AI-2-mediated signalling in bacteria. FEMS

microbiology reviews 37, 156-181.

Price, S.B., Cheng, C.M., Kaspar, C.W., Wright, J.C., DeGraves, F.J., Penfound, T.A., Castanie-Cornet, M.P.,

and Foster, J.W. (2000). Role of rpoS in acid resistance and fecal shedding of Escherichia coli O157:H7.

Applied and environmental microbiology 66, 632-637.

Ptashne, M. (1986). A genetic switch : gene control and phage [lambda] (Cambridge, Mass.

Palo Alto, Calif.: Cell Press ;

Blackwell Scientific Publications).

Rahmati, S., Yang, S., Davidson, A.L., and Zechiedrich, E.L. (2002). Control of the AcrAB multidrug efflux

pump by quorum-sensing regulator SdiA. Molecular microbiology 43, 677-685.

Raibaud, O., Vidal-Ingigliardi, D., and Kolb, A. (1991). Genetic studies on the promoter of malT, the gene

that encodes the activator of the Escherichia coli maltose regulon. Research in microbiology 142, 937-942.

Randall-Hazelbauer, L., and Schwartz, M. (1973). Isolation of the bacteriophage lambda receptor from

Escherichia coli. Journal of bacteriology 116, 1436-1446.

Rutherford, S.T., and Bassler, B.L. (2012). Bacterial quorum sensing: its role in virulence and possibilities for

its control. Cold Spring Harb Perspect Med 2.

83

Sabag-Daigle, A., Soares, J.A., Smith, J.N., Elmasry, M.E., and Ahmer, B.M. (2012). The acyl homoserine

lactone receptor, SdiA, of Escherichia coli and Salmonella enterica serovar Typhimurium does not respond

to indole. Applied and environmental microbiology 78, 5424-5431.

Samson, J.E., Magadan, A.H., Sabri, M., and Moineau, S. (2013). Revenge of the phages: defeating bacterial

defences. Nature reviews Microbiology 11, 675-687.

Saxena, S., and Gowrishankar, J. (2011). Modulation of Rho-dependent transcription termination in

Escherichia coli by the H-NS family of proteins. Journal of bacteriology 193, 3832-3841.

Sertic, V., and Boulgakov, N.A. (1936). Specific bacteriophages for flagellated bacterial varieties. Cr Soc Biol

123, 887-888.

Sharma, V.K., Bearson, S.M., and Bearson, B.L. (2010). Evaluation of the effects of sdiA, a luxR homologue,

on adherence and motility of Escherichia coli O157 : H7. Microbiology 156, 1303-1312.

Shimada, K., Ogasawara, H., Yamada, K., Shimura, M., Kori, A., Shimada, T., Yamanaka, Y., Yamamoto, K.,

and Ishihama, A. (2013). Screening of promoter-specific transcription factors: multiple regulators for the

sdiA gene involved in cell division control and quorum sensing. Microbiology 159, 2501-2512.

Shimada, T., Shimada, K., Matsui, M., Kitai, Y., Igarashi, J., Suga, H., and Ishihama, A. (2014). Roles of cell

division control factor SdiA: recognition of quorum sensing signals and modulation of transcription

regulation targets. Genes to cells : devoted to molecular & cellular mechanisms 19, 405-418.

Sitnikov, D.M., Schineller, J.B., and Baldwin, T.O. (1996). Control of cell division in Escherichia coli:

regulation of transcription of ftsQA involves both rpoS and SdiA-mediated autoinduction. Proceedings of

the National Academy of Sciences of the United States of America 93, 336-341.

Smith, J.L., Fratamico, P.M., and Yan, X. (2011). Eavesdropping by bacteria: the role of SdiA in Escherichia

coli and Salmonella enterica serovar Typhimurium quorum sensing. Foodborne pathogens and disease 8,

169-178.

Soares, J.A., and Ahmer, B.M. (2011). Detection of acyl-homoserine lactones by Escherichia and Salmonella.

Current opinion in microbiology 14, 188-193.

Sperandio, V., Torres, A.G., Giron, J.A., and Kaper, J.B. (2001). Quorum sensing is a global regulatory

mechanism in enterohemorrhagic Escherichia coli O157:H7. J Bacteriol 183, 5187-5197.

Starosta, A.L., Lassak, J., Jung, K., and Wilson, D.N. (2014). The bacterial translation stress response. FEMS

microbiology reviews.

Stoebel, D.M., Free, A., and Dorman, C.J. (2008). Anti-silencing: overcoming H-NS-mediated repression of

transcription in Gram-negative enteric bacteria. Microbiology 154, 2533-2545.

Takayanagi, Y., Tanaka, K., and Takahashi, H. (1994). Structure of the 5' upstream region and the regulation

of the rpoS gene of Escherichia coli. Molecular & general genetics : MGG 243, 525-531.

Tkachenko, A.G., Kashevarova, N.M., Karavaeva, E.A., and Shumkov, M.S. (2014). Putrescine controls the

formation of Escherichia coli persister cells tolerant to aminoglycoside netilmicin. FEMS microbiology

letters.

Tomasz, A. (1965). Control of the competent state in Pneumococcus by a hormone-like cell product: an

example for a new type of regulatory mechanism in bacteria. Nature 208, 155-159.

Turnbull, A.L., Kim, W., and Surette, M.G. (2012). Transcriptional regulation of sdiA by cAMP-receptor

protein, LeuO, and environmental signals in Salmonella enterica serovar Typhimurium. Canadian journal of

microbiology 58, 10-22.

Uchiyama, J., Nobue, Y., Zhao, H., Matsuzaki, H., Nagahama, H., Matsuoka, S., Matsumoto, K., and Hara, H.

(2010). Involvement of sigmaS accumulation in repression of the flhDC operon in acidic phospholipid-

deficient mutants of Escherichia coli. Microbiology 156, 1650-1660.

Van Houdt, R., Aertsen, A., Moons, P., Vanoirbeek, K., and Michiels, C.W. (2006). N-acyl-L-homoserine

lactone signal interception by Escherichia coli. FEMS microbiology letters 256, 83-89.

Veesler, D., and Cambillau, C. (2011). A common evolutionary origin for tailed-bacteriophage functional

modules and bacterial machineries. Microbiology and molecular biology reviews : MMBR 75, 423-433, first

page of table of contents.

84

Walters, M., and Sperandio, V. (2006). Quorum sensing in Escherichia coli and Salmonella. International

journal of medical microbiology : IJMM 296, 125-131.

Wang, M., Weiss, M., Simonovic, M., Haertinger, G., Schrimpf, S.P., Hengartner, M.O., and von Mering, C.

(2012). PaxDb, a database of protein abundance averages across all three domains of life. Molecular &

cellular proteomics : MCP 11, 492-500.

Wang, X.D., de Boer, P.A., and Rothfield, L.I. (1991). A factor that positively regulates cell division by

activating transcription of the major cluster of essential cell division genes of Escherichia coli. The EMBO

journal 10, 3363-3372.

Waterman, S.R., and Small, P.L. (2003). Transcriptional expression of Escherichia coli glutamate-dependent

acid resistance genes gadA and gadBC in an hns rpoS mutant. Journal of bacteriology 185, 4644-4647.

Wei, Y., Lee, J.M., Smulski, D.R., and LaRossa, R.A. (2001). Global impact of sdiA amplification revealed by

comprehensive gene expression profiling of Escherichia coli. Journal of bacteriology 183, 2265-2272.

Westra, E.R., Swarts, D.C., Staals, R.H., Jore, M.M., Brouns, S.J., and van der Oost, J. (2012). The CRISPRs,

they are a-changin': how prokaryotes generate adaptive immunity. Annual review of genetics 46, 311-339.

Xiao, H., Kalman, M., Ikehara, K., Zemel, S., Glaser, G., and Cashel, M. (1991). Residual guanosine 3',5'-

bispyrophosphate synthetic activity of relA null mutants can be eliminated by spoT null mutations. The

Journal of biological chemistry 266, 5980-5990.

Yamamoto, K., Yata, K., Fujita, N., and Ishihama, A. (2001). Novel mode of transcription regulation by SdiA,

an Escherichia coli homologue of the quorum-sensing regulator. Molecular microbiology 41, 1187-1198.

Yao, Y., Martinez-Yamout, M.A., Dickerson, T.J., Brogan, A.P., Wright, P.E., and Dyson, H.J. (2006). Structure

of the Escherichia coli quorum sensing protein SdiA: activation of the folding switch by acyl homoserine

lactones. Journal of molecular biology 355, 262-273.

Zhou, K., Zhou, L., Lim, Q., Zou, R., Stephanopoulos, G., and Too, H.P. (2011). Novel reference genes for

quantifying transcriptional responses of Escherichia coli to protein overexpression by quantitative PCR.

BMC molecular biology 12, 18.

Zhou, X., Meng, X., and Sun, B. (2008). An EAL domain protein and cyclic AMP contribute to the interaction

between the two quorum sensing systems in Escherichia coli. Cell research 18, 937-948.