The Indole Pulse: A New Pe spec i e on Indole Signalling
in
Esche ichia coli
Hannah Gaims e
1
*, Jehangi Cama
2
, Sil ia He na
´ndez-Ainsa
2
, Ul ich F. Keyse
2
, Da id K. Summe s
1
1Depa men o Gene ics, Uni e si y o Camb idge, Camb idge, Uni ed Kingdom, 2Ca endish Labo a o y, Uni e si y o Camb idge, Camb idge, Uni ed Kingdom
Abs ac
Indole has di e se signalling oles, including modula ion o bio ilm o ma ion, i ulence and s ess esponses. Changes a e
induced by indole concen a ions o 0.5–1.0 mM, simila o hose ound in he supe na an o Esche ichia coli s a iona y
phase cul u e. He e we desc ibe an al e na i e mode o indole signalling ha p omo es he su i al o E. coli cells du ing
long- e m s a iona y phase. A mu an ha has los he abili y o p oduce indole demons a es educed su i al unde hese
condi ions. Signi ican ly, he addi ion o 1 mM indole o he cul u e supe na an is insu icien o es o e long- e m su i al
o he mu an . We p o ide e idence ha he pe inen signal in his case is no 1 mM indole in he cul u e supe na an bu
a ansien pulse o in a-cellula indole a he ansi ion om exponen ial g ow h o s a iona y phase. Du ing his pulse he
cell-associa ed indole eaches a maximum o app oxima ely 60 mM. We a gue ha his is su icien o inhibi g ow h and
di ision by an ionopho e-based mechanism and causes he cells o en e s a iona y phase be o e esou ces a e exhaus ed.
The unused esou ces a e used o epai and main ain cells du ing he ex ended pe iod o s a a ion.
Ci a ion: Gaims e H, Cama J, He na
´ndez-Ainsa S, Keyse UF, Summe s DK (2014) The Indole Pulse: A New Pe spec i e on Indole Signalling in Esche ichia coli. PLoS
ONE 9(4): e93168. doi:10.1371/jou nal.pone.0093168
Edi o : Finba Hayes, Uni e si y o Manches e , Uni ed Kingdom
Recei ed No embe 14, 2013; Accep ed Feb ua y 28, 2014; Published Ap il 2, 2014
Copy igh : ß2014 Gaims e e al. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s
un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed.
Funding: HG was unded by a BBSRC DTG s uden ship, h p://www.bbs c.ac.uk/home/home.aspx, g an numbe PCAG-EJNF. JC, SHA and UFK acknowledge
suppo om an ERC s a ing g an , h p://e c.eu opa.eu/, g an numbe 261101 Passmemb ane. JC is also suppo ed by an In e nal G adua e S uden ship om
T ini y College, Camb idge, h p://www. in.cam.ac.uk/. The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o
he manusc ip .
Compe ing In e es s: The au ho s ha e decla ed ha no compe ing in e es s exis .
* E-mail: [email p o ec ed]
In oduc ion
Indole is a signalling molecule sec e ed by o e 85 species o
bac e ia including Esche ichia coli [1]. I is p oduced by he enzyme
yp ophanase (TnaA) ha con e s yp ophan in o indole,
py u a e, and ammonia [2]. E. coli makes e y li le indole du ing
exponen ial g ow h. Howe e yp ophanase exp ession is s ongly
up- egula ed by he s a iona y phase sigma ac o RpoS [3], so
indole p oduc ion ises as cells app oach s a iona y phase [4]. The
inal concen a ion o indole in s a iona y phase cul u e depends
upon he amoun o yp ophan in he g ow h medium. In LB
medium ee yp ophan can ange be ween 0.5–1 mM, and
supe na an concen a ions o indole also ypically each 0.5–
1 mM [5].
Indole has di e se signalling oles including modula ion o
bio ilm o ma ion, i ulence and s ess esponses [6–8]. I can also
ac as an in e kingdom signal and has been shown o a ec gene
exp ession in human en e ocy es [9] and memb ane po en ial in
mi ochond ia [10]. In each o hese cases he sys em is esponding
o ela i ely low (,1 mM), pe sis en concen a ions o indole.
Recen wo k has e ealed a ole o indole in plasmid s abili y.
Dime isa ion o he mul icopy plasmid ColE1 is a well-cha ac-
e ised cause o plasmid ins abili y [11]. The accumula ion o
plasmid dime s in a cell igge s he syn hesis o a 70n egula o y
RNA, Rcd, encoded wi hin he plasmid ce si e [12–13]. Rcd
exp ession leads o inhibi ion o cell di ision and i is sugges ed
ha he ac i a ion o his checkpoin gi es su icien ime o he
dime s o be con e ed o monome s by si e-speci ic ecombina-
ion a he plasmid ce si e [11].
The a ge o Rcd is yp ophanase [4] and Rcd binding educes
i s K
m
o yp ophan, s imula ing indole p oduc ion. Since he
addi ion o indole o he cul u e medium o g owing cells was
shown e e sibly o a es cell di ision [10], i was p oposed ha
indole p oduc ion by Rcd-ac i a ed yp ophanase was he
mechanism by which he Rcd checkpoin blocks cell di ision.
Indole has been shown o a es bac e ial g ow h and cell
di ision by ac ing as an ionopho e and making he cy oplasmic
memb ane pe meable o hyd ogen ions [10]. The consequen
loss o memb ane po en ial p e en s he unc ion o he
MinCDE sys em and localiza ion o he F sZ ing ha is
equi ed o cell di ision. Howe e , while his e ec is seen a
indole concen a ions o 3–5 mM, he maximum concen a ion
o indole in a cul u e supe na an cul u e is ypically 0.5–1 mM.
Unsu p isingly, he e o e, i has been unclea whe he he
e ec s o 3–5 mM indole upon bac e ial cells a e biologically
ele an .
He e we epo a no el mode o indole signalling du ing E. coli
s a iona y phase en y ha is impo an o iabili y in long- e m
s a iona y phase. We show ha du ing he ansi ion om
exponen ial o s a iona y phase cells ansien ly expe ience a e y
high (.50 mM) concen a ion o indole and ha his pulse is
necessa y o long- e m s a iona y phase iabili y. This i s
example o a ‘‘pulse signalling’’ mechanism expands he epe oi e
o indole e ec s on bac e ia o include ionopho e-media ed e ec s
on cell di ision and g ow h. We specula e ha i may p o e o be
o widesp ead signi icance.
PLOS ONE | www.plosone.o g 1 Ap il 2014 | Volume 9 | Issue 4 | e93168
Ma e ials and Me hods
S ains and cul u e condi ions
E. coli BW25113, W3110, BW25113 D naA (Kanamcyin
esis an : Km
R
) and W3110 D naA we e ob ained om he Keio
collec ion [14]. Cells we e cul u ed ou inely in Lu ia Be ani (LB)
medium a 37uC, wi h shaking a 120 pm. O e nigh cul u es
we e dilu ed o OD
600
= 0.05 and allowed o g ow o 2 hou s
be o e samples we e emo ed o subsequen assays. Whe e
equi ed, indole (dissol ed in e hanol) was added; e hanol alone
was added o con ols whe e app op ia e. To assess colony
o ming uni s (CFU), cul u es we e dilu ed app op ia ely and
sp ead on o LA pla es. These we e incuba ed a 37uC and colonies
coun ed he ollowing day.
Mixed cul u e expe imen
O e nigh cul u es o BW25113 and BW25113 D naA (Km
R
)
cells we e dilu ed in o esh LB medium, in iplica e, o an OD
600
o 0.05 a an ini ial a io o 99:1 (BW25113 cells o BW25113 D
naA). These cul u es we e g own a 37uC, wi h shaking, o
24 hou s. A 0, 2, 4, 6, 8, and 24 hou s he samples we e dilu ed
app op ia ely and we e pla ed on o bo h LA and LA +kanamycin
pla es. This allowed he p opo ion o D naA cells in he
popula ion o be de e mined.
Ko acs Assay
Ko acs assay is a commonly-used echnique in he ield which
gi es consis en esul s in di e en labo a o ies [4–5]. The cell
associa ed indole concen a ions p esen ed in his epo ( o a
maximum o 60 mM) a e much highe han cul u e medium
concen a ions measu ed p e iously using he Ko acs assay.
Howe e , hese a e calcula ed cell-associa ed alues. The ‘‘ aw’’
assay unde pinning he calcula ion is in he con en ional, linea
ange o he assay.
To assay indole in cul u e supe na an s, a sample (1 ml) om a
g owing cul u e was emo ed, he OD
600
measu ed and cells
ha es ed by cen i uga ion a 11337 x g o 15 seconds
(Eppendo Minispin mic o uge). The supe na an was emo ed
and assayed: 300 ml o Ko acs Reagen (10 g o p-dime hylamino-
benzaldehyde dissol ed in a mix u e o 50 ml o HCl and 150 ml
o amyl alcohol) was added o he supe na an and incuba ed o 2
minu es. A 50 ml po ion was emo ed and added o 1 ml o HCl-
amyl alcohol solu ion (75 ml o HCl and 225 ml o amyl alcohol).
The abso bance a 540 nm was measu ed (Gene Quan 1300, GE
Spec opho ome e ). The concen a ion o indole in he supe na-
an was calcula ed using a calib a ion cu e.
To assay indole wi hin he cell pelle , a modi ied me hod was
used. A sample (1 ml) om a g owing cul u e was emo ed, he
OD
600
measu ed and cells ha es ed by cen i uga ion a 11337 x
g o 15 seconds. The supe na an was disca ded and he cell pelle
assayed: 300 ml o Ko acs Reagen was added o he cell pelle o
2 minu es. This lysed he cells and allowed he Ko acs eagen o
eac wi h indole. This mix u e was pipe ed in o 1 ml LB be o e a
50 ml po ion was emo ed and added o 1 ml o HCl-amyl
alcohol solu ion. The abso bance a 540 nm was measu ed (Gene
Quan 1300, GE Spec opho ome e ). The concen a ion o indole
was calcula ed using a calib a ion cu e. The samples used o
gene a e he calib a ion cu e con ained an app op ia e densi y o
indole non-p oducing cells as well as indole a a known
concen a ion. In addi ion, he blank o he assay was gene a ed
om cul u e medium con aining cells bu no added indole.
The e o e, any backg ound due o cell deb is is aken in o accoun .
Using he concen a ion o indole ob ained om he calib a ion
cu e allowed us o de e mine he numbe o moles o indole
p esen in he cell pelle (Ip). Using he OD
600
o he cul u e, we
can calcula e he o al bac e ial cell olume (Vc) con ained in he
pelle [15] and hence he appa en cell associa ed indole
concen a ion (C
A
)as shown in equa ion 1.
CA~Ip
Vc
ð1Þ
Calcula ions o he appa en cell associa ed indole concen a-
ions based on cell pelle assays may o e es ima e he indole
concen a ion. This is due o he inclusion (in he Ip e m) o indole
dissol ed in he cul u e supe na an and apped in he cell pelle .
The calcula ion ou lined below allows us o co ec o his e o .
Vp~VczVs~Vcz1
2Vc~3
2Vcð2Þ
Equa ion 2 s a es ha ha he olume o he pelle is equal o
he sum o he olume o he apped supe na an and he olume
o he cells. Re e ence 16 s a es ha in a cell pelle , app oxima ely
1/3 o he pelle will co espond o apped supe na an .
Cp~Ip
Vp
~2
3|Ip
Vc
~2
3CAð3Þ
Cp~Cc|RVczCs|RVs~2
3Ccz1
3Csð4Þ
whe e C
A
is he appa en cell associa ed indole concen a ion in
he pelle , I
p
is he numbe o moles o indole in he pelle , V
c
is he
olume o he cells, V
p
is he olume o he pelle and V
s
is
he olume o he supe na an , C
p
is he concen a ion o indole in
he pelle , C
c
is he concen a ion o indole in he cells and C
s
is
he concen a ion o indole in he apped supe na an , R
Vp
is he
olume ac ion o he cells in he pelle (R
Vp
=2
3) and R
Vs
is he
olume ac ion o he supe na an in he pelle (R
Vs
=1
3), aken
om [16].
Fo example, whe e C
A
= 60 mM and C
s
= 4 mM, so by using
Equa ions 3 and 4 we can es ima e C
c
= 58 mM.
UV Abso bance Assay o Indole Concen a ion
Abso bance measu emen s we e ca ied ou using a Va ian
Ca y 300 Bio UV-VIS Spec opho ome e . Unless o he wise
s a ed, chemicals and media we e ob ained om Sigma Ald ich
UK. W3110 D naA cells we e cul u ed o e nigh a 37uC a 250
pm in shake lasks con aining 30 ml LB medium con aining
kanamycin (30 mgml
21
). Cells we e washed by cen i uging he
cul u e o 10 min a app oxima ely 2685 x g (4,000 pm in a
Ha ie 18/80 Re ige a ed Cen i uge), emo ing he supe na an
and esuspending he pelle in phospha e bu e ed saline (PBS;
p epa ed by dissol ing PBS able s in an app op ia e olume o
deionized (Milli-Q) wa e ; his was hen au ocla ed and s e ile
il e ed (0.22 mm, Millipo e) p io o use). A e washing he cells
we e again ha es ed by cen i uga ion and e-suspended in he
desi ed amoun o PBS. The OD
600
o he cul u es was de e mined
bo h be o e washing (in LB) and a e washing (in PBS).
Indole Pulse Signalling
PLOS ONE | www.plosone.o g 2 Ap il 2014 | Volume 9 | Issue 4 | e93168
One ml aliquo s o cells in PBS we e ans e ed o Eppendo
ubes (Axygen, 1.5 ml MaxyClea Mic o ubes). App op ia e
amoun s o 75 mM indole s ock solu ion (in absolu e e hanol)
we e added o ob ain he desi ed supe na an concen a ions. A
con ol wi hou indole was also p epa ed.
A e indole addi ion, he samples we e o exed and he cells
we e pelle ed by cen i uga ion o 5 minu es a 3500 g (6,000 pm
in a The mo Scien i ic He aeus F esco 17 Cen i uge). The
supe na an s ( e e ed o as ‘bac e ial supe na an s’) we e dilu ed
1:20 in PBS and hei abso bance alues measu ed in he 250–300
nm ange. The dilu ion was equi ed o he abso bance alues o
lie wi hin he linea ange o he spec opho ome e . The indole
peak abso bance a 268 nm was de e mined o all he samples.
Backg ound sub ac ion was applied using he con ol.
Sepa a ely, abso bance cu es we e measu ed in he 250–300
nm ange o indole solu ions (0–2.5 mM) p epa ed in PBS (a e
dilu ing 1:20 in PBS simila o he s eps desc ibed abo e). The
indole abso bance peak a 268 nm was measu ed, backg ound
sub ac ion was pe o med using he ‘no-indole’ sample and he
peak abso bance alues we e used o plo a calib a ion cu e o
Abso bance s Ex e nal Indole Concen a ion.
Using he abo e calib a ion cu e and he abso bance alue
measu ed om he bac e ial supe na an , he indole concen a ion
o he bac e ial supe na an was de e mined. Sub ac ing his om
he known ex e nal concen a ion allows us o de e mine he
numbe o moles o indole p esen in he bac e ial pelle . Using he
OD
600
ob ained (in PBS), we can calcula e he o al bac e ial cell
olume con ained in he pelle [15] and hence he appa en cell
associa ed indole concen a ion.
De e mina ion o he E. coli lipids-bu e pa i ion
coe icien o indole
E.coli o al lipid ex ac (A an i Pola Lipids) dissol ed in
chlo o o m was added o a 5 ml ial. The o ganic sol en was
emo ed by high- acuum d ying o 3–4 hou s o gi e a 5 mg
lipid ilm. Solu ion A (5 mM indole, Sigma Ald ich, in 15 mM
phospha e bu e (PB; pH = 7)) was p epa ed by adding he
app op ia e olume o a s ock solu ion o indole in e hanol o he
15 mM PB bu e . The lipid ilm was dispe sed by adding 3.5 ml
o solu ion A and sonica ing using an ul asonic ba h (G an ,
MXB6, 160W) o a ound 2 hou s a 25uC. In o de o sepa a e
he lipid and aqueous phases, samples we e cen i uged wi h a
Beckman Coul e -Op ima max XP Ul acen i uge a 50,000 x g
o 45 min a 25uC. A e ul acen i uga ion, lipids appea ed as a
pelle a he bo om o he cen i uga ion ubes. The aqueous
phase was aken ca e ully wi h a Pas eu pipe e and kep in an
Eppendo ube o he spec opho ome ic measu emen s.
The E. coli lipids-bu e pa i ion coe icien o indole was
de e mined using a Va ian Ca y 300 Bio UV-VIS Spec opho-
ome e . A calib a ion cu e o abso bance s indole concen a ion
a 268 nm was ini ially ob ained by p epa ing indole solu ions in
he 15 mM PB bu e in a ange o concen a ions o 0–0.2 mM
(linea ange). The indole concen a ion o he aqueous phase was
de e mined by measu ing he abso p ion a 268 nm, no malising
he alues wi h espec o he ‘no-indole’ and using he calib a ion
cu e o es ima e he concen a ion. No-indole samples con aining
he same quan i y o lipids (5 mg) dispe sed in PB bu e bu
lacking indole, we e submi ed o he same ul acen i uga ion and
sepa a ion p ocess. The concen a ion o indole in he lipid phase
was calcula ed by sub ac ing he amoun o indole in he aqueous
phase om he o al amoun p esen in he ini ial solu ion A.
The pa i ion coe icien (P) was hen calcula ed using equa ion
[5] [17] whe e C
i
is he concen a ion o indole in solu ion A, C
w
is
he concen a ion o indole in he aqueous phase, w
w
is he weigh
o he aqueous phase (3500 mg) and w
l
he weigh o he lipid
phase (5 mg) in he samples.
P~Ci{Cw
Cw
|ww
wl
ð5Þ
Es ima ion o he olume o he lipid memb ane in E. coli
The olume o he lipid memb ane in E.coli (0.04 mm
3
) has been
de e mined using equa ion [6] and applying simple geome ical
calcula ions. In hese calcula ions, E. coli is assumed o ha e a
cylind ical shape. lis he leng h o E.coli (1.6 mm) [15], is he
es ima ed adius o he cylinde (0.55 mm) and dis he hickness o
he lipid bilaye (a ound 5nm) [18].
Volume lipid memb ane~2pl ðÞz2p 2

|dð6Þ
The olume o he en i e E. coli is 1.5 mm
3
, so he lipid
memb ane ep esen s 1/40
h
o he o al cell olume.
Es ima ion o he molecula a io indole:lipid and he
weigh pe cen age indole:lipid
The es ima ion o he molecula a io indole:lipid (0.460.1) has
been made using Equa ion [8] (simpli ied om Equa ion [7])
whe e dis he densi y o he lipids (1000 g L
21
), V
lipid
is he
olume o he lipid memb ane in a single E. coli (0.04 mm
3
), M
lipid
is he molecula weigh o E. coli o al ex ac lipids (811.5 g
mol
21
), N
A
is he A ogad o cons an (6.023610
23
), C
indole
is he
concen a ion o indole added o he supe na an (0.005 mol L
21
)
and P is he E. coli lipids-bu e pa i ion coe icien (92.9624.5).
Numbe indole molecules
Numbe lipid molecules ~P|Cindole|Vlipid |NA
d|Vlipid |1
Mlipid

NA
ð7Þ
Numbe indole molecules
Numbe lipid molecules ~P|Cindole|Mlipid
dð8Þ
The pe cen age in weigh indole:lipid has been calcula ed
mul iplying Equa ion [4] by he a io o molecula weigh s o
indole:lipid, namely (117.2:811.5).
Resul s
Indole p oduc ion inc eases apidly du ing en y in o
s a iona y phase
The concen a ion o indole in he supe na an o an L-b o h
cul u e o E. coli BW25113 was measu ed using he Ko acs assay
(Fig. 1). Consis en wi h epo s elsewhe e [5], he supe na an
concen a ion eached a maximum o 0.7–0.8 mM in s a iona y
phase. The concen a ion was low in la e exponen ial phase bu
ose apidly when he cul u e eached an op ical densi y (OD
600
)
Indole Pulse Signalling
PLOS ONE | www.plosone.o g 3 Ap il 2014 | Volume 9 | Issue 4 | e93168
o 1.0–1.5. This co esponded o he pe iod o ansi ion be ween
exponen ial g ow h and s a iona y phase.
The apid accumula ion o indole in he supe na an (5- old
inc ease o e 30 min) e lec ed an inc eased p oduc ion a e pe
cell a he han simply an inc ease in cell numbe s, since he
OD
600
o he cul u e inc eased only 1.4- old du ing his pe iod.
Indole p oduc ion is equi ed o long- e m s a iona y
phase iabili y
In o de o assess he unc ional signi icance o indole
p oduc ion du ing he onse o s a iona y phase, we compa ed
he g ow h and iabili y o BW25113 and BW25113 D naA (a
yp ophanase knock-ou ha p oduces no indole) o e 10 days a
37uC. Fo he i s h ee days he densi y o BW25113 cul u es was
signi ican ly lowe han BW25113 D naA, implying ha indole
p oduc ion inhibi s g ow h du ing s a iona y phase en y. How-
e e o e he subsequen 7 days he OD
600
o he mu an cul u e
declined, while he densi y o he wild- ype con inued slowly o
inc ease (Fig. 2A). Consequen ly, a e 10 days in s a iona y phase,
i was he wild- ype cul u e ha had he highe op ical densi y.
Op ical densi y measu emen s p o ide an es ima e o he
numbe o cells in a cul u e bu no whe he hese cells a e ali e
o dead. To compa e he iabili ies o cells in TnaA
+
and TnaA
2
cul u es, samples we e dilu ed and sp ead on L-aga pla es o
es ima e iable coun s (Fig. 2B). We obse ed ha o e 9 days in
s a iona y phase he colony o ming uni s (CFU) in he mu an
cul u e dec eased by app oxima ely 75%, compa ed o he wild-
ype cul u e whe e CFU dec eased by less han 30%. One possible
in e p e a ion o hese da a is ha indole p oduc ion causes g ow h
o he wild- ype cul u e o slow be o e esou ces a e exhaus ed and
he a ailabili y o unused esou ce assis s long- e m su i al in
s a iona y phase.
In an a emp o es o e he s a iona y phase iabili y o he
D naA mu an s ain, indole (1 mM) was added o he cul u e
medium when he OD
600
eached 1.5. This mimicked he apid
accumula ion o indole in he cul u e supe na an o wild- ype cells
a his poin . Howe e , he addi ion o 1 mM indole had no e ec
on ei he he cul u e densi y o he long- e m iabili y o he
mu an (Fig. 2).
The ailu e o 1 mM indole o es o e he long- e m iabili y o
he D naA mu an was ini ially su p ising. Howe e , indole
p oduc ion by wild- ype cells du ing s a iona y phase en y
appea s o slow hei g ow h (Fig. 2A), and indole concen a ions
o below 2–3 mM a e known o ha e li le o no e ec on g ow h
[4]. A signi ican ly highe (4–5 mM) indole concen a ion is
equi ed o inhibi g ow h o cell di ision. Because indole
p oduc ion du ing s a iona y phase en y is e y apid, i is
possible ha a highe concen a ion o indole exis s inside he cells
du ing he p oduc ion pe iod. I a su icien ly high concen a ion
we e eached, his migh explain he slowing o g ow h obse ed
o he wild- ype s ain.
To es he p oposal ha i is he in e nal a he han he
ex e nal concen a ion o indole which de e mines he kine ics o
en y in o s a iona y phase, a mixed cul u e expe imen was
pe o med. The cul u e con ained bo h wild- ype and D naA
mu an cells, a an ini ial a io 99:1. This allowed us o obse e he
beha iou o D naA cells in he p esence o le els o indole ha
occu in he wild- ype cul u e supe na an . The e is a clea
p edic ion ha i he kine ics o s a iona y phase en y a e a ec ed
by he ex e nal le el o indole, hen all he cells in he mixed
cul u e should en e s a iona y phase a he same ime and he
p opo ion o he D naA mu an cells should emain cons an .
Howe e i an ele a ed in e nal le el o indole in he wild- ype
p oduce cells egula es s a iona y phase en y hen he wild- ype
cells will en e s a iona y phase ea lie and he D naA mu an cells
will inc ease hei p opo ion in he popula ion.
Samples we e aken om he mixed cul u e a egula in e als
and he p opo ion o mu an cells was measu ed. The p opo ion
inc eased om 0.760.4% when he cul u es we e se up, o
4.460.2% a 24 hou s. This co esponds o a 6- old inc ease in
he p opo ion o mu an cells in 24 hou s. Fu he mo e, he
g ea es inc ease in he p opo ion o mu an cells (a 4- old
inc ease) occu ed be ween 6 and 8 hou s and was he e o e a e
indole was p oduced by he wild- ype cells. This esul sugges s
ha esou ces unused by wild- ype cells emained in he
supe na an whe e hey we e a ailable o use by he mu an s.
Cell associa ed indole inc eases apidly be o e s a iona y
phase
To es he possibili y ha a high in acellula indole concen-
a ion exis s du ing s a iona y phase en y, we used he Ko acs
assay o measu e appa en cell-associa ed indole. Cul u e samples
we e cen i uged, he supe na an emo ed and he cell pelle lysed
by he addi ion o Ko acs eagen . The concen a ion o indole in
he lysa e was measu ed by he Ko acs assay and was con e ed o
he appa en cell-associa ed concen a ion using an es ima e o E.
coli cell olume [15]. We use he e m ‘appa en ’ o acknowledge
ha he concen a ion has no , a his s age, been co ec ed o any
con ibu ion om indole in he cul u e supe na an ha is apped
in he cell pelle . The esul showed ha appa en cell-associa ed
indole inc eased sha ply bu ansien ly du ing s a iona y phase
en y, eaching a maximum o 60 mM (Fig. 3). Also signi ican was
he obse a ion ha he appa en cell-associa ed indole concen-
a ion ne e d opped below 15 mM and was always subs an ially
highe han he supe na an concen a ion (Fig. 1).
The e ec s o de ined concen a ions o indole added o he
cul u e medium on he g ow h and di ision o E. coli a e well
es ablished bu his is no ue o cell-associa ed indole. To help
Figu e 1. Indole is p oduced apidly o e a 30 minu e pe iod
du ing he onse o s a iona y phase. A cul u e o g owing
BW25113 cells was sampled egula ly. The OD
600
was measu ed and he
samples we e cen i uged o emo e cells, and he supe na an was
assayed o indole using Ko acs assay. Da a shown a e he mean alues
6s anda d de ia ion o h ee independen epea s.
doi:10.1371/jou nal.pone.0093168.g001
Indole Pulse Signalling
PLOS ONE | www.plosone.o g 4 Ap il 2014 | Volume 9 | Issue 4 | e93168
in e p e he da a o Fig. 3, known concen a ions o indole (0–
5 mM) we e added o s a iona y phase cul u es o a D naA mu an
s ain. Cells we e ha es ed and he pelle assayed o appa en
cell-associa ed indole (Fig. 4). A supe na an concen a ion o
0.75 mM ( ypical o s a iona y phase cul u es o wild- ype cells)
ga e an appa en cell-associa ed concen a ion o app ox. 20 mM,
while an appa en cell associa ed concen a ion o 60 mM ( he
maximum de ec ed du ing s a iona y phase en y) esul ed om a
supe na an concen a ion o app ox. 4 mM.
The appa en cell associa ed indole concen a ion can be
co ec ed o he con ibu ion o indole om he cul u e
supe na an . I has p e iously been es ima ed ha a cell pelle is
likely o con ain app oxima ely 30% cul u e supe na an in
addi ion o he cells [16]. We ha e desc ibed in he ma e ial and
me hods a calcula ion o compensa e o he con ibu ion o he
ex acellula indole in he pelle . In Fig. 4, we see ha 4 mM
indole in he cul u e medium ga e an appa en cell associa ed
indole o 60 mM. Applying he co ec ion, he ac ual cell
associa ed indole concen a ion is shown o be 58 mM a he
han 60 mM. Thus, he con ibu ion o indole om apped
supe na an o ou appa en cell associa ed indole concen a ions
is negligible and well wi hin he expe imen al e o s o ou assay.
To con i m ha he high concen a ion o cell-associa ed indole
was no an a e ac o he Ko acs assay, he ela ionship be ween
cell-associa ed and supe na an indole was in es iga ed by an
independen echnique. Indole non-p oducing (TnaA
2
) cells we e
suspended in medium con aining a known concen a ion o indole
and hen ha es ed by cen i uga ion. The amoun o indole
emo ed om he medium by he cells was es ima ed by using an
UV abso bance assay o measu e he indole concen a ion in he
medium be o e and a e ha es ing (see Ma e ials and Me hods
o de ails). When he indole concen a ion in he medium was
1 mM, a cell-associa ed concen a ion o 763 mM was calcula ed
(8 epea s). This compa es wi h a alue o 1763 mM when
cell-associa ed indole was measu ed using he Ko acs assay (Fig.
4). The di e ence be ween he wo alues may be a sys ema ic
e o due o p ocedu al di e ences be ween he assays, howe e
hey a e s ill wi hin an o de o magni ude o each o he and hey
help con i m ha he high cell associa ed alues a e no simply an
a e ac o he Ko acs assay.
Indole dis ibu ion wi hin cells
Appa en cell-associa ed indole is an a e age o he indole
concen a ions in he cy oplasm and in he lipid memb anes. To
de e mine he ela i e indole a ini y o hese wo cell componen s
we measu ed he pa i ion coe icien (P) o indole be ween wa e
and E. coli o al lipid. We ob ained a alue o log (P) = 1.9560.12
(n = 9), which is simila o he calcula ed alue o 2.17 o oc anol-
wa e pa i ioning epo ed in [19]. Thus indole has an app ox.
90- old highe a ini y o lipid han wa e , due o he hyd opho-
bici y o he a oma ic ing o he molecule.
Discussion
The e ec s o low (0.5 –1.0 mM) concen a ions o indole on
bac e ial physiology a e well es ablished [1]. In addi ion, ecen
s udies ha e shown ha a highe (4–5 mM) concen a ions, indole
can egula e bac e ial g ow h and di ision h ough i s ac ion as a
p o on ionopho e [10]. Howe e , he biological ele ance o hese
e ec s has been open o ques ion since concen a ions abo e
1 mM a e no seen in cul u e supe na an s. In his epo we
desc ibe a mode o indole signalling ha is independen o he
supe na an concen a ion, bu is d i en by a ansien , high
concen a ion o indole inside he cell.
A cul u e o wild- ype E. coli makes he ansi ion om
exponen ial o s a iona y phase ea lie han a cul u e o an indole
non-p oducing mu an . The wild- ype also displays highe iabili y
in long- e m s a iona y phase, and a likely explana ion is ha
Figu e 2. Non indole p oducing mu an s ini ially g ow o a highe densi y and a e mo e iable han indole p oducing coun e pa s,
bu a e signi ican ly less iable in he long e m. The densi y (OD
600
: A) and iabili y (CFU 610
26
/ ml :B) o wild- ype, mu an and mu an wi h
1 mM indole added we e assessed o e 10 days. Da a shown a e he mean alues 6s anda d de ia ion o h ee independen epea s.
doi:10.1371/jou nal.pone.0093168.g002
Indole Pulse Signalling
PLOS ONE | www.plosone.o g 5 Ap il 2014 | Volume 9 | Issue 4 | e93168

esou ces unused o g ow h emain a ailable o he wild- ype o
epai and main enance du ing he pe iod o s a a ion as
sugges ed by he mixed cul u e esul . We we e unable o es o e
wild- ype beha iou o he mu an by supplemen ing he cul u e
medium wi h 1 mM indole. This is pe haps unsu p ising when
one emembe s ha 1 mM indole has no de ec able e ec upon
ei he g ow h o di ision o E. coli [4]. The indole supplemen a ion
expe imen was designed so ha indole was added o he cul u e a
he same ime, and wi h he same kine ics, as i appea s in wild-
ype cul u e (Fig. 1). The only di e ence be ween he indole-
supplemen ed mu an cells and wild- ype cells was ha he wild-
ype cells expe ienced a ansien cell-associa ed indole concen-
a ion o 60 mM (Fig. 3). We he e o e conclude ha he
subsequen di e ence in g ow h and iabili y o he wo cul u es
was due o he p esence o absence o his pulse o cell-associa ed
indole. This dis inc ion be ween he e ec s o ex e nal and cell-
associa ed indole is u he demons a ed in he mixed cul u e
expe imen . He e, mu an cells ha did no expe ience he indole
pulse en e ed s a iona y phase la e han co-cul u ed wild- ype
cells.
The 6- old inc ease in indole non-p oducing cells in he mixed
cul u e expe imen also sugges s ha esou ces unused by wild-
ype cells emain in he cul u e medium and a e a ailable o use
by o he bac e ia. This seems a li le su p ising since he ese e is
hen a ailable o use by ‘‘chea s’’ ha con inue o g ow un il he
esou ce is exhaus ed. I his is ue, we would p edic ha du ing
long- e m s a iona y phase he mu an cells would main ain hei
ele a ed p opo ion in he mixed cul u e bu ha bo h mu an and
wild- ype cells would die a he highe a e exhibi ed by mu an
cells in pu e cul u e.
The pulse is well de ined and las s app ox. 20 min (Fig. 3). We
sugges ha his co esponds o a pe iod when indole is being
p oduced as e han i can escape om he cell by di usion
h ough he memb ane. The s a o he pulse can hen be
explained by he RpoS-induced up- egula ion o yp ophanase
exp ession as s a iona y phase app oaches [3]. The end o he
pulse is mo e di icul o explain since yp ophanase is p esumably
s ill p esen . Howe e he o al amoun o indole p oduced by a
cul u e is limi ed by he amoun o ee yp ophan in he g ow h
medium [5], so he end o he pulse p obably e lec s he ime
when all a ailable yp ophan has been con e ed o indole. I is
impo an o no e ha he pulse du a ion o 20 min is an a e age
o all cells in he cul u e. I he e is signi ican popula ion
he e ogenei y in he ime o pulsing, he pulse du a ion in
indi idual cells may be sho e and he maximum concen a ion
highe han he alues we p esen he e.
In e p e a ion o ou da a is complica ed by he ac ha E. coli
cells, by i ue o hei lipid memb anes (and possibly o he cell
componen s), ha e a g ea e a ini y o indole han hei aqueous
su oundings. This is why cell-associa ed concen a ions a e
always highe han supe na an concen a ions, no jus du ing
he indole pulse. Since in p e ious epo s he e ec s o indole on
he g ow h and di ision o E. coli ha e been de e mined
expe imen ally by he addi ion o indole o he g ow h medium
(4, 20), i is help ul o con e appa en cell-associa ed indole o
‘‘supe na an -equi alen concen a ions’’ using he da a o Fig. 4.
Thus when he appa en cell-associa ed concen a ion is 20 mM in
la e s a iona y phase, he supe na an -equi alen concen a ion is
app ox. 0.75 mM and his is insu icien o ha e any e ec on
g ow h. Howe e , when he appa en cell-associa ed concen a-
ion is 60 mM a he peak o he pulse, he supe na an -equi alen
concen a ion is app oxima ely 4 mM (Fig. 4) and high enough o
cause signi ican inhibi ion o g ow h [4]. Thus ou da a sugges
ha he magni ude o he indole pulse du ing s a iona y phase
en y is su icien o egula e g ow h and di ision by he ionopho e
mechanism desc ibed by Chime el e al. [10]. I has p e iously
been epo ed ha a he onse o s a iona y phase, E. coli cells
become smalle [21] due o con inued elonga ion and di ision.
Figu e 3. Appa en cell associa ed indole accumula es apidly
a he onse o s a iona y phase. A cul u e o g owing BW25113
cells was sampled egula ly The OD
600
was measu ed and he samples
we e cen i uged o pelle cells. The esul an cell pelle was assayed o
indole using Ko acs assay. Da a shown a e he mean alues 6s anda d
de ia ion o h ee independen epea s.
doi:10.1371/jou nal.pone.0093168.g003
Figu e 4. The ela ionship be ween appa en cell associa ed
concen a ions o indole and supe na an concen a ions o
indole. Known concen a ions o indole we e added o he supe na-
an o s a iona y phase BW25113 D naA cells, he mix u e was
o exed and cen i uged, and he esul an pelle assayed o indole
using Ko acs assay. Do ed lines indica es he appa en cell associa ed
indole a he peak o he pulse (60 mM) and he co esponding
supe na an concen a ion and indica es he appa en cell associa ed
indole when cells a e in s a iona y phase and he co esponding
supe na an concen a ion (0.75 mM). Da a shown a e he mean alues
6s anda d de ia ion o a leas h ee independen epea s.
doi:10.1371/jou nal.pone.0093168.g004
Indole Pulse Signalling
PLOS ONE | www.plosone.o g 6 Ap il 2014 | Volume 9 | Issue 4 | e93168
Cell size depends on he ela i e con ibu ions o cell g ow h and
cell di ision. Indole has been shown o a ec bo h hese p ocesses
and i g ow h is inhibi ed ea lie , o mo e s ongly, han cell
di ision by he indole pulse hen smalle cells would esul .
I is clea om ou da a ha he indole pulse is a non-
equilib ium phenomenon. A he peak o he pulse, he
concen a ion o appa en cell-associa ed indole (60 mM; Fig. 3)
is 150- old highe han he supe na an concen a ion (0.4 mM;
Fig. 1). As he sys em e u ns owa ds equilib ium in s a iona y
phase, he appa en cell-associa ed concen a ion (20 mM) exceeds
he supe na an concen a ion (0.75 mM) by less han 30- old.
The indole:wa e pa i ion coe icien (log (P) = 1.95) indica es
ha he equilib ium concen a ion o indole in he memb ane i sel
(as opposed o he o al cell-associa ed concen a ion) will also be
subs an ially highe han he concen a ion in he su ounding
medium. I he lipid memb ane accoun s o 1/40
h
o he o al cell
olume, he e will be app oxima ely wice as much indole in he
memb anes as in he cy oplasm. We es ima e he mola
indole:lipid a io as 0.4 (60.1):1 when 5 mM indole is added o
he g ow h medium (see Ma e ials and Me hods o de ails).
Howe e , he molecula weigh o E. coli lipids is conside ably
g ea e han he molecula weigh o indole, so i is use ul also o
exp ess he a io o indole:lipid as 0.05:1 by mass. Ei he way, we
en isage ha he ema kably high concen a ion o indole in he
lipid memb ane du ing he pulse may ha e a signi ican di ec
e ec on he p ope ies o memb ane p o eins, in addi ion o i s
e ec on ionic pe meabili y.
Does he a ini y o E. coli lipids o indole explain comple ely
he high cell-associa ed indole concen a ions ha we measu ed
unde equilib ium condi ions (Fig. 4)? To answe his we need o
assume ha all cell componen s, o he han lipid, ha e he same
a ini y o indole as wa e . We can hen calcula e he expec ed
cell-associa ed concen a ion o any gi en ex e nal indole
concen a ion, based on he lipid:wa e pa i ion coe icien and
he p opo ion o a cell ha is lipid. This calcula ion p edic s ha
he cell-associa ed concen a ion should be app oxima ely 3 imes
highe han he concen a ion in he su ounding medium. This is
signi ican ly di e en om he expe imen al esul s shown in Fig.
4, whe e he cell-associa ed concen a ion is app oxima ely 15
imes highe han he concen a ion in he g ow h medium. This
seems o imply ha some addi ional componen in he cell may
ha e a high a ini y o indole. Al e na i ely, i is possible ha he
pa i ion coe icien measu ed in i o may no en i ely e lec he
si ua ion in i o whe e he lipid composi ion may a y and he lipid
solubili y o indole may be in luenced by he memb ane cha ge.
Au ho Con ibu ions
Concei ed and designed he expe imen s: HG JC SH-A UFK DKS.
Pe o med he expe imen s: HG JC SH-A. Analyzed he da a: HG JC SH-
A UFK DKS. Con ibu ed eagen s/ma e ials/analysis ools: HG JC SH-
A. W o e he pape : HG JC SH-A UFK DKS.
Re e ences
1. Lee JH, Lee J (2010) Indole as an in e cellula signal in mic obial communi ies.
FEMS Mic obiology Re iews 34: 426–444.
2. New on WA, Snell EE (1962) An inducible yp ophan syn he ase in yp ophan
auxo ophs o Esche ichia coli. P oceedings o he Na ional Academy o Sciences o
he Uni ed S a es o Ame ica 48: 1431–1433.
3. Lacou S, Landini P (2004) Sigma(S)-dependen gene exp ession a he onse o
s a iona y phase in Esche ichia coli: Func ion o sigma(S)-dependen genes and
iden i ica ion o hei p omo e sequences. Jou nal o Bac e iology 186: 7186–
7195.
4. Chan EL, Summe s DK (2007) Indole signalling con ibu es o he s able
main enance o Esche ichia coli mul icopy plasmids. Molecula Mic obiology 63:
35–43.
5. Li G, Young KD (2013) Indole p oduc ion by he yp ophanase TnaA in
Esche ichia coli is de e mined by he amoun o exogenous yp ophan.
Mic obiology-Sgm 159: 402–410.
6. Hi akawa H, Hayashi-Nishino M, Yamaguchi A, Nishino K (2010) Indole
enhances acid esis ance in Esche ichia coli. Mic obial Pa hogenesis 49: 90–94.
7. Di Ma ino P, Fu sy R, B e L, Sunda a aju B, Phillips RS (2003) Indole can ac
as an ex acellula signal o egula e bio ilm o ma ion o Esche ichia coli and o he
indole-p oducing bac e ia. Canadian Jou nal o Mic obiology 49: 443–449.
8. Hi akawa H, Inazumi Y, Masaki T, Hi a a T, Yamaguchi A (2005) Indole
induces he exp ession o mul id ug expo e genes in Esche ichia coli. Molecula
Mic obiology 55: 1113–1126.
9. Bansal T, Alaniz RC, Wood TK, Jaya aman A (2010) The bac e ial signal
indole inc eases epi helial-cell igh -junc ion esis ance and a enua es indica o s
o in lamma ion. P oceedings o he Na ional Academy o Sciences o he Uni ed
S a es o Ame ica 107: 228–233.
10. Chime el C, Field CM, Pine o-Fe nandez S, Keyse UF, Summe s DK (2012)
Indole p e en s Esche ichia coli cell di ision by modula ing memb ane po en ial.
Biochimica E Biophysica Ac a-Biomemb anes 1818.
11. Summe s DK, She a DJ (1984) Mul ime iza ion o high copy numbe
plasmids causes ins abili y- ColE1 encodes a de e minan essen ial o plasmid
monome iza ion and s abili y. Cell 36: 1097–1103.
12. Pa ien ME, Summe s DK (1993) ColE1 mul ime o ma ion igge s inhibi ion
o Esche ichia coli cell di ision Molecula Mic obiology 9: 1089–1095.
13. Balding C, Blaby I, Summe s D (2006) A mu a ional analysis o he ColE1-
encoded cell cycle egula o Rcd con i ms i s ole in plasmid s abili y. Plasmid
56: 68–73.
14. Da senko KA, Wanne BL (2000) One-s ep inac i a ion o ch omosomal genes
in Esche ichia coli K-12 using PCR p oduc s. P oceedings o he Na ional
Academy o Sciences o he Uni ed S a es o Ame ica 97: 6640–6645.
15. Volkme B, Heinemann M (2011) Condi ion-Dependen Cell Volume and
Concen a ion o Esche ichia coli o Facili a e Da a Con e sion o Sys ems
Biology Modeling. Plos One 6.
16. Reid DF, F ank HA (1966) Iso opic me hod o es ima ing mic obial cell
olumes. Jou nal o Bac e iology 92: 639– 644.
17. Vazquez JL, Me ino S, Domenech O, Be langa M, Vinas M, e al. (2001)
De e mina ion o he pa i ion coe icien s o a homologous se ies o
cip o loxacin: in luence o he N-4 pipe azinyl alkyla ion on he an imic obial
ac i i y. In e na ional Jou nal o Pha maceu ics 220: 53–62.
18. Albe s B (2005) Molecula Biology o he Cell: Ga land Science.
19. Kama aju K, Smi h J, Wang J, Roy V, Sin im HO, e al. (2011) E ec s on
Memb ane La e al P essu e Sugges Pe mea ion Mechanisms o Bac e ial
Quo um Signaling Molecules. Biochemis y 50: 6983–6993.
20. Pine o-Fe nandez S, Chime el C, Keyse UF, Summe s DK (2011) Indole
T anspo ac oss Esche ichia coli Memb anes. Jou nal o Bac e iology 193: 1793–
1798.
21. Ake lund T, No ds om K, Be nande R (1995) Analysis o cell-size and DNA
con en in exponen ially g owing and s a iona y-phase ba ch cul u es o
Esche ichia coli. Jou nal o Bac e iology 177: 6791–6797.
Indole Pulse Signalling
PLOS ONE | www.plosone.o g 7 Ap il 2014 | Volume 9 | Issue 4 | e93168