How Bugs Kill Bugs: Progress and Challenges in Bacteriocin Research
Colicin translocation across the Escherichia coli
outer membrane
Nicholas G. Housden and Colin Kleanthous1
Department of Biochemistry, University of Oxford, Oxford OX1 3QU, U.K.
Abstract
We are investigating how protein bacteriocins import their toxic payload across the Gram-negative cell
envelope, both as a means of understanding the translocation process itself and as a means of probing
the organization of the cell envelope and the function of the protein machines within it. Our work focuses
on the import mechanism of the group A endonuclease (DNase) colicin ColE9 into Escherichia coli, where
we combine in vivo observations with structural, biochemical and biophysical approaches to dissect the
molecular mechanism of colicin entry. ColE9 assembles a multiprotein ‘translocon’ complex at the E. coli
outer membrane that triggers entry of the toxin across the outer membrane and the simultaneous jettisoning
of its tightly bound immunity protein, Im9, in a step that is dependent on the protonmotive force. In the
present paper, we focus on recent work where we have uncovered how ColE9 assembles its translocon
complex, including isolation of the complex, and how this leads to subversion of a signal intrinsic to the
Tol–Pal assembly within the periplasm and inner membrane. In this way, the externally located ColE9 is able
to ‘connect’ to the inner membrane protonmotive force via a network of protein–protein interactions that
spans the entirety of the E. coli cell envelope to drive dissociation of Im9 and initiate entry of the colicin
into the cell.
Introduction
The outer membrane of Gram-negative bacteria such
as Escherichia coli provides a formidable barrier against
antibiotics, digestive enzymes, detergents and the host
immune system, aiding survival in hostile environments [1].
Nutrient uptake across the outer membrane is facilitated
by numerous proteins providing either active or passive
transport routes for nutrients into the periplasm. In order
for colicins to unleash their potent cytotoxic activities, they
must first breach their target cell and reach the appropriate
location within the cell for toxin deployment. In the case of
pore-forming colicins, this means crossing the E. coli outer
membrane so that a depolarizing pore can be inserted into the
inner membrane. Enzymatic nuclease colicins face the more
formidable task of crossing the outer and inner membranes to
reach their cytoplasmic targets. The passage of colicins across
the E. coli cell envelope requires the hijacking of multiple
target cell proteins within the outer membrane, periplasm and
inner membrane, with intrinsic disorder playing a pivotal role
in the multiple protein–protein interactions that occur. The
present review focuses on recent progress in understanding
the multiple protein–protein interactions involved in the
translocation of group A enzymatic colicins, such as Col
(colicin) E9, across the E. coli cell envelope.
Key words: colicin, Escherichia coli, native disorder, OmpF, translocation..
Abbreviations used: Col, colicin; FM, fluorescein-5-maleimide; FRET, fluorescence resonance
energy transfer; ITC, isothermal titration calorimetry; IUTD, intrinsically unstructured translocation
domain; OBS, OmpF-binding site; TAMRA, 5(6)-carboxytetramethylrhodamine; TBE, TolB-binding
epitope; TMR, tetramethylrhodamine-6-maleimide.
1
To whom correspondence should be addressed (email [email protected]).
Biochem. Soc. Trans. (2012) 40, 1475–1479; doi:10.1042/BST20120255
The N-terminal translocation and central receptor-binding
domains of the DNase (ColE2, ColE7, ColE8 and ColE9),
the rRNase (ColE3, ColE4 and ColE6) and tRNase (ColE5)
colicins are highly conserved, reflecting a common delivery
apparatus for multiple different killing activities. Therefore
understanding the translocation pathway for any of these
toxins will shed light on the common route of access,
potentially allowing this to be exploited for the delivery of
novel antimicrobial agents.
Figure 1 highlights the key host proteins within the E. coli
cell envelope which are hijacked to form the ColE9 translocon
required for the entry of the C-terminal cytotoxic domain
into the cytoplasm.
Receptor binding
To ensure adequate uptake of scarce nutrients such as vitamin
B12 and iron siderophores, specific receptors exist in the outer
membrane for their active transport into the periplasm. One
such protein is BtuB, the E. coli vitamin B12 receptor, which
forms a 22-stranded β-barrel within the outer membrane,
the lumen of which is plugged by a 136 residue N-terminal
globular plug domain [2]. Within the N-terminus of this plug
domain, there is a TonB box, which, upon ligand binding,
undergoes a conformational change such that it is presented
to TonB. TonB forms a complex with ExbB and ExbD
and is coupled to the inner membrane protonmotive force,
providing the energy source for nutrient uptake [3].
The first step towards colicin translocation into the target
cell is the formation of high-affinity complex between its
receptor-binding domain and BtuB. Unlike vitamin B12
C The
C 2012 Biochemical Society
Authors Journal compilation 1475
1476
Biochemical Society Transactions (2012) Volume 40, part 6
Figure 1 Overview of the interactions of enzymatic group A colicins at the E. coli cell envelope
Colicin is concentrated on to the cell surface through a high-affinity interaction with BtuB, allowing OBS1 (shown in red)
located in the IUTD to bind within the OmpF pore. Upon dissociation of OBS1, the IUTD travels through the outer membrane
allowing OBS2 (shown in orange) to bind OmpF, presenting the TBE to the periplasm where it can bind TolB. Colicin binding
promotes the interaction of TolB with TolA, giving the colicin access to the inner membrane potential to drive entry of the
cytotoxic domain into the cytoplasm. Figure assembled from PDB files 1UJW (BtuB–Col E3 R-domain), 1JCH (ColE3–Im3), 3O0E
(OmpF), 2W8B (TolB–Pal) and 2IVZ (TolB–TBE).
uptake, group A enzymatic colicins show no requirement
for TonB, despite binding to a TonB-dependent receptor. The
in vitro interaction of ColE9 with detergent-solubilized BtuB
has been characterized biophysically by ITC (isothermal
titration calorimetry), where binding was found to be in the
low nanomolar range [4]. Rigidification of ColE9 through
the introduction of an engineered disulfide bond across the
top of the R-domain [5] had little impact on the observed
thermodynamics of complex formation, suggesting that no
major conformational changes occur within the colicin upon
complex formation. X-ray crystal structures of the ColE2
and ColE3 receptor-binding domains in complex with BtuB
have been published [6,7] showing the apex of the R-domain
to interact with the extracellular loops and the top of the
plug domain, which remains in place obscuring the lumen
of the β-barrel. Although interaction with BtuB does not
provide a route of passage across the outer membrane, it
serves to concentrate the colicin on to the surface of its target
bacterium allowing the N-terminal 83 amino acids of the
IUTD (intrinsically unstructured translocation domain) to
locate and bind the OmpF in the outer membrane.
Porin recruitment
The E. coli outer membrane is rich in the highly abundant
porins (∼105 /cell) OmpF and OmpC which allow the
passive diffusion of small (<600 Da) hydrophilic molecules
across the outer membrane. These proteins are trimeric,
assembled from 16-stranded β-barrel monomer subunits,
each of which forms a channel running across the outer
membrane. The expression levels of OmpF, OmpC and PhoE
are controlled by environmental conditions, with PhoE only
being produced under conditions of phosphate starvation.
C The
C 2012 Biochemical Society
Authors Journal compilation Whereas the role of trimeric porins such as OmpF, OmpC
and PhoE in the activity of group A enzymatic colicin activity
has long been established [8], evidence for a direct interaction
between the ColE9 and OmpF was first presented in 2005 [4],
with the extraction of intact BtuB–ColE9–Im9His6 –OmpF
complex assembled at the cell surface and captured by nickelaffinity chromatography. The ability of ColE9–Im9His6 added
to cells expressing OmpF and BtuB to form a BtuB–ColE9–
Im9His6 –OmpF complex requires the ColE9 IUTD; in the
absence of these 83 disordered residues, only the BtuB–ColE9
complex is observed.
Truncations of the IUTD revealed two OBSs (OmpFbinding sites): one within the extreme N-terminus and a
second beyond the TBE (TolB-binding epitope). ITC studies
on the binding of OmpF to a series of truncated IUTDs
allowed the two sites to be mapped to linear peptides T2–18
(OBS1, S2 GGDGRGHNTGAHSTSG18 ) and T54–63 (OBS2,
I54 HWGGGSGRG63 ), binding OmpF with affinities of 2 μM
and 24 μM respectively in 20 mM potassium phosphate
buffer (pH 6.5) at 25◦ C [9]. Both OBS1 and OBS2 bind
OmpF with a stoichiometry of three peptides per OmpF
trimer with both peptides competing for a common site of
interaction on OmpF. Electrostatics are likely to play an
important role in the interactions of both peptides which have
an overall positive charge with the mildly cation-selective
OmpF.
Two X-ray crystal structures of OmpF bound to T1–83
[10] and T2–18 [9] reveal density within the OmpF lumen
corresponding to the bound peptide, indicating that this is
the route of passage of the IUTD across the outer membrane
(Figure 2). With a high-resolution structure of the OmpF–
OBS1 complex remaining elusive, details of the interactions
between the peptide and OmpF lumen are unclear and the
How Bugs Kill Bugs: Progress and Challenges in Bacteriocin Research
Figure 2 Structure of the OmpF–ColE9 OBS1 peptide complex
(a) OmpF monomer shown in green ribbon with OBS1 peptide (residues 2–16) modelled into electron density within the
lumen. (b) Overlay of T1–83 ColE3 peptide backbone (yellow) [10] and T2–18 ColE9 peptide (red) [9] shown within the lumen
of OmpF (cutaway surface shown in grey). Reproduced from Housden, N.G., Wojdyla, J.A., Korczynska, J., Grishkovskaya,
I., Kirkpatrick, N., Brzozowski, A.M. and Kleanthous, C. (2010) Directed epitope delivery across the Escherichia coli outer
c 2010 National
membrane through the porin OmpF. Proc. Natl. Acad. Sci. U.S.A. 107, 21412–21417 with permission. Academy of Science.
directionality of the peptide within the OmpF lumen is
somewhat ambiguous. Alanine-scanning mutagenesis across
OBS1 reveals that, whereas several side chains have a minor
impact upon binding, there are no ‘hotspots’ critical to
complex formation. OBS1 is glycine-rich (35%), and mutation of any of these residues to proline abrogates binding,
whereas mutagenesis to alanine significantly affects
binding, suggesting flexibility to be key in the threading of
the peptide into the OmpF lumen [9].
The physiological relevance of these binding epitopes is
highlighted through assays of colicin cytotoxicity where the
loss of either region impairs colicin cytotoxicity, whereas
the combined loss of both epitopes inactivates the colicin. A
sequential binding model in which OBS1 binds OmpF in an
electrostatically driven interaction, followed by dissociation
via the periplasmic face of OmpF and the engagement of
OBS2 in its place, results in the passage of the intervening
TBE, which shows no binding with OmpF, across the outer
membrane.
Complex formation between ColE9 and
TolB
The Tol system comprises TolQ, TolR and TolA in the inner
membrane, TolB within the periplasm and Pal anchored in the
inner leaflet of the outer membrane is ubiquitous in Gramnegative bacteria and is known to function in maintaining the
integrity of the outer membrane [11,12]. Although the Tol
system is not essential, deletion of any of the tol genes results
in a characteristic phenotype where periplasmic contents
leak into the extracellular environment, the outer membrane
forms blebs or ruffles, and the cell becomes sensitive to
large antibiotics and detergents such as SDS, which would
normally be excluded from the cell.
TolQ and TolR are homologues of ExbB and ExbD of
the Ton system, and are coupled to the protonmotive force
[13,14]. By parasitizing TolB in the periplasm, which in
turn binds TolA of the TolQRA complex within the inner
membrane, the colicin may indirectly access the energy
potential of the inner membrane providing a driving force
for the internalization of the colicin.
The ColE9 five-residue TolB box (D35 GSGW39 ), critical
to TolB binding, which was identified previously by sitedirected mutagenesis [15,16], has been extended to the
16-residue TBE (G32 ASDGSGWSSENNPWG47 ) through
cross-linking studies and ITC measurements [17]. The
thermodynamic parameters of TBE binding to TolB are
comparable with those seen for full-length ColE9 (K d ∼1 μM
at 20◦ C in 50 mM potassium phosphate, pH 7.5, 50 mM NaCl
and 10 mM EDTA). Binding is characterized by a favourable
enthalpy change (H ∼ − 15 kcal/mol; 1 kcal = 4.184 kJ)
and an unfavourable entropy change (TS ∼ − 7 kcal/mol)
consistent with the disordered peptide becoming ordered
upon complex formation. With the TBE and the full-length
ColE9 showing similar thermodynamics upon binding TolB,
the disorder-to-order transition within the IUTD is likely to
be restricted to within the TBE.
A number of X-ray crystal structures have been
determined for TolB alone [18,19], in complex with Pal [20]
and in complex with the TBE peptide [17]. The TolB–Pal and
C The
C 2012 Biochemical Society
Authors Journal compilation 1477
1478
Biochemical Society Transactions (2012) Volume 40, part 6
TolB–TBE structures confirm the prediction from binding
studies, cross-linking and site-directed mutagenesis that the
Pal and TBE sites are overlapping [20]. Despite the TolB–
TBE complex having only half the buried surface area of the
TolB–Pal complex, the TBE adopts a conformation allowing
it to mimic key interactions of the TolB–Pal complex.
Kinetic basis of ColE9–TBE complex
formation
Under equilibrium conditions in the presence of EDTA, the
K d values of the Pal–TolB and TBE–TolB complexes are
∼50 nM and ∼1 μM respectively [17], making it difficult to
envisage how the colicin could displace Pal from TolB. In
the presence of Ca2 + , the affinity of the Pal–TolB complex
is weakened and the TBE–TolB complex enhanced such that
both complexes form with an affinity of ∼90 nM.
Recent fluorescence studies using engineered FRET
(fluorescence resonance energy transfer) pairs across the Pal–
TolB and TBE–TolB complexes revealed that competitive
recruitment of TolB by ColE9 is kinetically driven [21].
Sites for the positioning of FRET pairs within the complexes
were selected such that labelling would not affect binding
as verified through ITC measurements on the wild-type
and labelled mutant complexes. Pre-equilibrium fluorescence
measurements of the Pal–TolB complex were made using
FM (fluorescein-5-maleimide)-labelled S139C Pal binding
TMR (tetramethylrhodamine-6-maleimide)-labelled Q280C
TolB with complex formation being monitored through the
quench in the FM donor emissions. Equivalent measurements
for the TBE–TolB complex were made with TAMRA
[5(6)-carboxytetramethylrhodamine]- labelled TBE peptide
binding FM-labelled Q280C TolB, with TAMRA acceptor
emissions being used to monitor complex formation. Both the
TBE–TolB and the Pal–TolB complexes were found to form
via single-step binding reactions. In the absence of Ca2 + , the
association rate of the Pal–TolB and TBE–TolB complexes
were 7.6×104 M − 1 ·s − 1 and 1.9×105 M − 1 ·s − 1 respectively,
whereas, in the presence of Ca2 + , the rates were 5.4×104
M − 1 ·s − 1 and 2.2×105 M − 1 ·s − 1 respectively. With the TBE
binding TolB 2.5-fold faster than Pal in the absence of Ca2 +
and 4-fold faster in the presence of Ca2 + the TBE–TolB
complex will be favoured if all three proteins are present in
equal abundance.
Formation of the ColE9–TolB–TolA complex
Of all of the TolB crystal structures, the N-terminus is only
apparent in the TolB–Pal complex. Conformational changes
induced in TolB by Pal binding reveal a canyon on the
surface of TolB into which the N-terminus binds, forming an
antiparallel β-sheet. The self-association of the N-terminus
with the body of TolB results in 1700 Å2 (1 Å = 0.1 nm) of
buried surface area. Deletion of all, or part of this N-terminal
sequence results in a tol phenotype, showing it to be essential
to the function of the Tol system. The TolA–TolB in vitro
C The
C 2012 Biochemical Society
Authors Journal compilation complex can be detected through chemical cross-linking, but
the interaction is lost upon full or partial deletion of the TolB
N-terminus [22].
Although weak, the TolA–TolB complex can be analysed
by ITC, revealing an entropically driven binding interaction
dependent upon the N-terminus of TolB with a K d of
∼40 μM. The heats of binding are lost when the titration
is performed in the presence of Pal, whereas binding becomes
exothermic with a K d of ∼13 μM in the presence of ColE9.
Binding of Pal and ColE9 appear to have antagonistic effects
on the ability of TolB to bind TolA, with Pal binding making
the TolA-binding region of TolB unavailable, whereas ColE9
presents the epitope ready for TolA binding.
Conclusions
In order for enzymatic group A colicins to gain entry to
their target cell, they must subvert many proteins within
the target cell. Intrinsic disorder plays key roles in the
interactions of ColE9 with OmpF and TolB and is also key
to the resulting TolB–TolA interaction. The IUTD highlights
known features of intrinsically unstructured proteins, namely
displaying multiple linear epitopes in a relatively small
sequence of protein and the ability to ‘flycast’ to find binding
partners. The OBS1–OmpF crystal structure highlights an
underappreciated property of intrinsic disorder and that is
the ability of a linear epitope to pass through a small pore,
in this case allowing another binding epitope to pass across a
membrane to reach its binding partner.
Funding
C.K. is funded by the Wellcome Trust [grant number 082045] and
the Biotechnology and Biological Sciences Research Council [grant
number BB/G020671/02].
References
1 Nikaido, H. (2003) Molecular basis of bacterial outer membrane
permeability revisited. Microbiol. Mol. Biol. Rev. 67, 593–656
2 Chimento, D.P., Mohanty, A.K., Kadner, R.J. and Wiener, M.C. (2003)
Substrate-induced transmembrane signaling in the cobalamin
transporter BtuB. Nat. Struct. Biol. 10, 394–401
3 Postle, K. and Kadner, R.J. (2003) Touch and go: tying TonB to transport.
Mol. Microbiol. 49, 869–882
4 Housden, N.G., Loftus, S.R., Moore, G.R., James, R. and Kleanthous, C.
(2005) Cell entry mechanism of enzymatic bacterial colicins: porin
recruitment and the thermodynamics of receptor binding. Proc. Natl.
Acad. Sci. U.S.A. 102, 13849–13854
5 Penfold, C.N., Healy, B., Housden, N.G., Boetzel, R., Vankemmelbeke, M.,
Moore, G.R., Kleanthous, C. and James, R. (2004) Flexibility in the
receptor-binding domain of the enzymatic colicin E9 is required for
toxicity against Escherichia coli cells. J. Bacteriol. 186, 4520–4527
6 Kurisu, G., Zakharov, S.D., Zhalnina, M.V., Bano, S., Eroukova, V.Y.,
Rokitskaya, T.I., Antonenko, Y.N., Wiener, M.C. and Cramer, W.A. (2003)
The structure of BtuB with bound colicin E3 R-domain implies a
translocon. Nat. Struct. Biol. 10, 948–954
7 Sharma, O., Yamashita, E., Zhalnina, M.V., Zakharov, S.D., Datsenko, K.A.,
Wanner, B.L. and Cramer, W.A. (2007) Structure of the complex of the
colicin E2 R-domain and its BtuB receptor: the outer membrane colicin
translocon. J. Biol. Chem. 282, 23163–23170
How Bugs Kill Bugs: Progress and Challenges in Bacteriocin Research
8 Mock, M. and Pugsley, A.P. (1982) The BtuB group col plasmids and
homology between the colicins they encode. J. Bacteriol. 150,
1069–1076
9 Housden, N.G., Wojdyla, J.A., Korczynska, J., Grishkovskaya, I., Kirkpatrick,
N., Brzozowski, A.M. and Kleanthous, C. (2010) Directed epitope delivery
across the Escherichia coli outer membrane through the porin OmpF.
Proc. Natl. Acad. Sci. U.S.A. 107, 21412–21417
10 Yamashita, E., Zhalnina, M.V., Zakharov, S.D., Sharma, O. and Cramer,
W.A. (2008) Crystal structures of the OmpF porin: function in a colicin
translocon. EMBO J. 27, 2171–2180
11 Lazzaroni, J.C., Fognini-Lefebvre, N. and Portalier, R. (1989) Cloning of
the excC and excD genes involved in the release of periplasmic proteins
by Escherichia coli K12. Mol. Gen. Genet. 218, 460–464
12 Webster, R.E. (1991) The tol gene products and the import of
macromolecules into Escherichia coli. Mol. Microbiol. 5, 1005–1011
13 Cascales, E., Lloubes, R. and Sturgis, J.N. (2001) The TolQ–TolR proteins
energize TolA and share homologies with the flagellar motor proteins
MotA–MotB. Mol. Microbiol. 42, 795–807
14 Kampfenkel, K. and Braun, V. (1993) Membrane topologies of the TolQ
and TolR proteins of Escherichia coli: inactivation of TolQ by a missense
mutation in the proposed first transmembrane segment. J. Bacteriol.
175, 4485–4491
15 Bouveret, E., Rigal, A., Lazdunski, C. and Benedetti, H. (1998) Distinct
regions of the colicin A translocation domain are involved in the
interaction with TolA and TolB proteins upon import into Escherichia coli.
Mol. Microbiol. 27, 143–157
16 Garinot-Schneider, C., Penfold, C.N., Moore, G.R., Kleanthous, C. and
James, R. (1997) Identification of residues in the putative TolA box which
are essential for the toxicity of the endonuclease toxin colicin E9.
Microbiology 143, 2931–2938
17 Loftus, S.R., Walker, D., Mate, M.J., Bonsor, D.A., James, R., Moore, G.R.
and Kleanthous, C. (2006) Competitive recruitment of the periplasmic
translocation portal TolB by a natively disordered domain of colicin E9.
Proc. Natl. Acad. Sci. U.S.A. 103, 12353–12358
18 Abergel, C., Bouveret, E., Claverie, J.M., Brown, K., Rigal, A., Lazdunski, C.
and Benedetti, H. (1999) Structure of the Escherichia coli TolB protein
determined by MAD methods at 1.95 Å resolution. Structure 7,
1291–1300
19 Carr, S., Penfold, C.N., Bamford, V., James, R. and Hemmings, A.M.
(2000) The structure of TolB, an essential component of the
tol-dependent translocation system, and its protein–protein interaction
with the translocation domain of colicin E9. Structure 8, 57–66
20 Bonsor, D.A., Grishkovskaya, I., Dodson, E.J. and Kleanthous, C. (2007)
Molecular mimicry enables competitive recruitment by a natively
disordered protein. J. Am. Chem. Soc. 129, 4800–4807
21 Papadakos, G., Housden, N.G., Lilly, K.J., Kaminska, R. and Kleanthous, C.
(2012) Kinetic basis for the competitive recruitment of TolB by the
intrinsically disordered translocation domain of colicin E9. J. Mol. Biol.
418, 269–280
22 Bonsor, D.A., Hecht, O., Vankemmelbeke, M., Sharma, A., Krachler, A.M.,
Housden, N.G., Lilly, K.J., James, R., Moore, G.R. and Kleanthous, C.
(2009) Allosteric β-propeller signalling in TolB and its manipulation by
translocating colicins. EMBO J. 28, 2846–2857
Received 25 September 2012
doi:10.1042/BST20120255
C The
C 2012 Biochemical Society
Authors Journal compilation 1479