Sugar-Coating Bacteria with
Lipopolysaccharides
Bacterial lipopolysaccharide assembly requires more than 50 gene
products, including several recently identified factors
Andrew C. McCandlish and Thomas J. Silhavy
ram-negative bacteria produce li- may lead to severe inflammation, vasodilation,
G
popolysaccharide (LPS), the endo- and potentially fatal septic shock.
toxin that causes septic shock and is For the bacteria, LPS is the major lipid molec-
a leading cause of death in critically ular species in the outer leaflet of the outer
ill patients. Because LPS is the pre- membrane (OM). LPS molecules interact with
dominant surface-exposed lipid in Gram-nega- one another, forming a tight matrix along the
tive bacteria, it is one of the first compounds to surface of the cell, preventing free diffusion of
interact with host cells. Once LPS binds the compounds to and from the environment. Ow-
Toll-like receptor-4 (TLR4) complex on mam- ing to this barrier function, LPS is essential for
malian host endothelial cells, a signaling cascade viability for nearly all gram-negative bacteria.
stimulates the innate immune response, recruit- Because the bacterial outer membrane is assem-
ing macrophage cells and releasing cytokines at bled without an obvious biochemical energy
the site of infection, and also starting the adap- source (the periplasm is devoid of adenosine
tive immune response. Although these responses triphosphate, ATP), understanding LPS biogen-
generally benefit the host, a high bacterial load esis has proved a fascinating problem for us and
other bacterial cell biologists. Also spurring
us on is the possibility that, by better under-
standing the factors involved in assembling
Summary the LPS matrix, we will identify novel tar-
gets for antimicrobial therapies.
· Lipopolysaccharide (LPS) molecules interact in
the outer membrane (OM) of gram-negative Andrew C. McCan-
bacteria, preventing free diffusion of com- dlish is a recent
pounds from and into such cells. Ph.D. graduate of
Challenges Facing Bacteria in
· The SecYEG complex translocates both -bar- the Silhavy Lab and
rel proteins and lipoproteins across the inner Assembling Outer Membrane
is currently the
membrane (IM) from the bacterial cytoplasm Components
Manager of Strate-
into the periplasm, where such proteins are
shuttled to separate OM assembly sites. Gram-negative bacteria such as Escherichia gic Planning and
coli are surrounded by two membrane bi- Business Develop-
· Either soluble proteins within the periplasm
shuttle LPS from the IM to a docking site at the layers. The OM, which serves as a perme- ment at Flexible
OM, or LPS travels through points of contact ability barrier to the outside milieu, is asym- Medical Systems,
between the IM and OM. metric. The characteristic lipid species of the Princeton, N.J., and
· LPS is assembled at the cell surface either by a outer leaflet of the OM is almost exclusively Thomas J. Silhavy
two-step process in which such molecules are LPS, whereas the inner leaflet of the OM is the Warner Lam-
first delivered to the inner leaflet of the OM and and both leaflets of the inner membrane bert-Parke Davis
then flipped to the outer leaflet, or by a con- (IM) of such cells are composed of phospho- Professor of Molec-
certed process in which LPS molecules are in-
lipids. The OM also contains lipoproteins, ular Biology, Prince-
serted directly into the outer leaflet.
-barrel proteins such as porins which al- ton University,
low selective passage of small, water-solu- Princeton, N.J.
Volume 2, Number 6, 2007 / Microbe Y 289
FIGURE 1 obvious source of biochemical energy
such as ATP.
How Proteins Move
to the Bacterial OM
Proteins destined for the outer mem-
brane (OMPs) are synthesized in the
cytoplasm with a short amino-terminal
signal sequence containing about 25
amino acids (Fig. 1). Specialized pro-
teins, designated SecA and SecB, bind
these precursor proteins and direct
them to the SecYEG channel in the in-
ner membrane, through which they
cross the IM, ultimately ending up in
the periplasm. During this energy-
dependent process, a peptidase along
the periplasmic face of the IM cleaves
the signal sequence on each translo-
cated OM protein. Lipoproteins con-
tain lipid moieties that anchor them in
the membrane after the signal sequence
Lipoprotein and -barrel protein (OMP) biogenesis. Lipoproteins and OMPs are synthe- is removed.
sized in the cytoplasm with an amino-terminal signal peptide. SecA and SecB bind the Lipoproteins that are destined for the
nascent protein and direct it to the Sec machinery where it is translocated across the OM, such as Braun's lipopoprotein
inner membrane. The signal sequence is cleaved by signal peptidases. Lipoproteins are
modified by the addition of lipids which anchor them in the membrane. LolCDE facilitates (LPP), interact with an IM complex,
the release of OM lipoproteins from the IM, while IM lipoproteins have a "Lol-avoidance" designated LolCDE, that uses energy
signal and thus remain in the IM. LolA shuttles the lipoprotein across the periplasm to derived from ATP molecules to extract
LolB where it is inserted into the inner leaflet of the OM. Several periplasmic factors are
known to be involved OMP biogenesis, but their precise role (ie. as a shuttle analogous the lipoprotein from the IM. IM-des-
to LolA) is unclear. A complex consisting of the essential OMP YaeT and four associated tined lipoproteins have a "Lol-avoid-
lipoproteins facilitates OMP insertion into the OM. ance" signal (typically, an aspartate res-
idue at the 2 position) that prevents
ble molecules in and out of the cell, and or- them from interacting with LolCDE,
ganelles such as pili. Immediately inside the OM thereby retaining them in the IM. The periplas-
is the periplasm, an aqueous compartment that mic chaperone protein LolA picks up OM-
is densely packed with proteins, including bind- bound lipoproteins from LolCDE and shuttles
ing proteins that function in nutrient transport them across the periplasm to an essential OM
and degradative enzymes such as alkaline phos- lipoprotein, LolB, where they are inserted into
phatase. the inner leaflet of the OM.
The inner membrane of gram-negative bacte- Signal peptidase releases -barrel OM pro-
ria is symmetric and, in addition to phospholip- teins into the periplasm, where chaperone pro-
ids, contains proteins involved in many func- teins such as SurA, Skp, and DegP prevent their
tions, including energy production, active misfolding and aggregating while delivering
transport, and signal transduction. The IM sep- them to the OM. Recently, our group identified
arates the periplasm from the innermost com- an essential OM complex that is required for
partment, the cytoplasm. The cytoplasm con- assembling outer membrane proteins. This com-
tains the E. coli chromosome and is where most plex contains the -barrel protein YaeT and
cellular components are synthesized. Molecules four lipoproteins.
destined for the OM must cross the IM and A common theme emerges from our knowl-
periplasm before inserting into the OM in their edge of -barrel and lipoprotein targeting and
appropriate conformations. During the final assembly in E. coli. In both pathways, the Sec-
stages of this process, they do so without an YEG complex translocates such proteins from
290 Y Microbe / Volume 2, Number 6, 2007
the cytoplasm. Following release from FIGURE 2
the IM, periplasmic proteins (LolA in
the case of lipoproteins and either SurA,
Skp, or DegP for -barrel proteins),
shuttle the OM-destined proteins
across the periplasm to an OM assem-
bly site, which is LolB for lipoproteins
and the YaeT complex for -barrel pro-
teins, where they are inserted into the
OM. Thus, although many details re-
main to be elucidated, we have some
conception of how proteins move from
the bacterial cytoplasm to the OM.
Two Competing Models To Explain
How LPS Moves to the OM
Our understanding of how LPS mole-
cules move into place in the OM is less
well developed. LPS molecules consist
of three distinct domains: lipid A,
core oligosaccharide, and O-antigen.
Though the precise structures of these
domains vary, there are common fea-
tures to each. For example, lipid A, Current model of LPS biogenesis. LPS is synthesized on the inner face of the IM then
which remains anchored inside the flipped to the outer leaflet of the IM by MsbA. As the IM-associated protein LptB and the
OM, generally consists of two diacyl- periplasmic protein LptA are known to be involved in LPS biogenesis, it is assumed they
play a role in shuttling LPS from the IM to the OM, but their precise function remains
glucosamine residues that typically are unknown. The gene encoding YrbK is located immediately upstream of the lptAB operon,
substituted with six hydrocarbon suggesting a possible role in LPS assembly, but its function is unknown. The Imp/RlpB
chains. Those two diacylglucosamines complex resides in the OM and targets LPS to its final destination in the outer leaflet of
the OM. Though the precise mechanism of Imp/RlpB is not known, two possibilities are
link to residues of 3-deoxy-D-manno- depicted here. In one case, LPS is inserted into the inner leaflet of the OM then flipped
oct-2-ulosonic acid (KDO) to serve as a to the outer leaflet. In the other case, LPS is received from the periplasm and inserted
scaffold, upon which the highly variable directly into the outer leaflet. Either function could be performed by Imp and/or RlpB.
Several lines of evidence suggest that LPS travels through zones of adhesion between
core oligosaccharide domain assembles. the IM and OM. The factors involved in LPS biogenesis could be independent entities
There are at least five different con- associated with a larger lipid or protein "Bayer's Bridge," or they could form a
served core structures in E. coli alone. periplasm-spanning complex as shown on the right.
In E. coli K-12, for example, the core
region contains four heptose, three glu-
cose, and one galactose residue. Additional sug- Researchers tend to agree on the early steps of
ars are added to form the O-antigen, a hyper- this process. For instance, following synthesis,
variable polysaccharide chain. Because laboratory LPS is flipped to the outer leaflet of the IM in an
strains of E. coli have lost their capacity to synthe- ATP-dependent step mediated by the essential
size the O-antigen, we use the term LPS to refer to IM protein, MsbA (Fig. 2). MsbA is a member of
the lipid A-core motif unless otherwise noted. the ABC-transporter family of proteins and
Most steps in assembling the lipid A-core LPS shares homology with mammalian multidrug
structure occur at the inner face of the IM, transporters. In E. coli strains containing tem-
presenting LPS with the same challenge as OM perature-sensitive msbA alleles, LPS accumu-
proteins in reaching the OM. Thus LPS some- lates in the inner leaflet of the IM at nonpermis-
how must flip from the inner to the outer side of sive temperatures, appearing in electron
the IM, cross the periplasmic space, and insert micrographs as invaginations of membranous
from the inner into the outer side of the OM. Is material into the cytoplasm. Such msbA mu-
the pathway for LPS biogenesis analogous to tants also display altered membrane density,
that of proteins? There are differing schools of while LPS can be isolated in IM fractions using
thought on some of these details. biochemical methods.
Volume 2, Number 6, 2007 / Microbe Y 291
FIGURE 3
Models for Imp/RlpB function. If Imp/RlpB flips LPS from the inner leaflet of the OM to the outer leaflet, then depleting Imp/RlpB will cause
accumulation of LPS in the OM (left panel). If Imp/RlpB receives LPS and inserts it directly into its final destination at the cell surface, then
LPS will accumulate in the IM upon Imp/RlpB depletion (at the last completed step in biogenesis, right panel). Upon Imp/RlpB depletion,
"extra" membrane material accumulates in the periplasm as visualized in electron micrographs. When MsbA function is disrupted, LPS
accumulates in the inner leaflet of the IM, and membranes invaginate into the cytoplasm. Thus, by determining whether the "extra"
membranes originate from the OM or the IM in Imp/RlpB depleted cells, we could differentiate between the proposed models of Imp/RlpB
action. The bottom panels depict this strategy diagrammatically.
However, there are two competing models for In any case, newly synthesized LPS appears at
how LPS is transported from the IM to the OM. discrete points along the bacterial cell surface.
According to one, LPS travels through points of When radiolabeling is used, LPS appears tran-
contact between the IM and OM (Fig. 2, right). siently in membranous fractions of novel den-
The validity of such zones of adhesion, which sity, according to Larry Rothfield of the
were identified by Manfred Bayer about 40 University of Connecticut Health Center in
years ago and came to be called Bayer's bridges, Farmington, Conn., and his collaborators, and
remains controversial. They appear at several these may be the adhesion zones seen in electron
points along the envelope, where the IM and micrographs. Moreover, according to Jan Tom-
OM appear to contact or come very close to one massen at the University of Utrecht in Utrecht,
another--at least, when depicted in thin-section the Netherlands, and his collaborators, radiola-
electron micrographs. However, some research- beled LPS reaches the OM even when ordinary
ers continue to argue that these contact sites are cells are converted into spheroplasts, draining
artifacts that disappear when cryofixation tech- those cells of their periplasmic contents. Adhe-
niques are used instead of chemical fixation sion zones could be what provides the structural
methods to visualize bacterial cellular struc- continuity in spheroplasts that enables LPS to
tures, an idea first suggested by the late Eduard move between the IM and OM.
Kellenberger. Meanwhile, according to the alternative
292 Y Microbe / Volume 2, Number 6, 2007
model, soluble proteins within the FIGURE 4
periplasm shuttle LPS from MsbA in
the IM to a docking site along the OM
(Fig. 2, left). This model closely paral-
lels what we know about the targeting
of -barrel proteins and lipoproteins to
the OM. Further and also consistent
with this view, two genes, lptB and
lptA, apparently play key roles in a
periplasmic intermediate-based path-
way for transporting LPS to the OM,
according to Alessandra Polissi of Uni-
versita degli Studi in Milano, Italy, and
`
her collaborators.
Through a large-scale genomic mu-
tagenesis approach, the Italian re-
searchers determined that these two
genes, which are located near KDO syn-
thesis genes along the E. coli chromo-
some, are essential for E. coli viability.
LptB protein appears to be a member of
the ABC-transporter family, while
LptA is periplasmic, according to
Polissi and her collaborators. Cells de-
pleted of either gene product cannot
properly assemble LPS. In particular, Electron micrographs of Imp- and RlpB-depleted cells. The top panels show the "extra"
membrane material in the periplasm of cells depleted of Imp or RlpB. These cells have
these strains accumulate misassembled higher levels of LPS than wild-type cells. Presumably the "extra" membranes originate
LPS in the IM. at the site of LPS accumulation as described in Fig. 3; that is, there is such a high level
These two models may not be mutu- of local LPS accumulation in the OM or IM that LPS-rich membranes are protruding into
the periplasm from the site of accumulation. The bottom panels present high-resolution
ally exclusive. Periplasmic factors may electron micrographs initiated to determine whether the membrane protrusions originate
transport LPS across the periplasm from the IM or the OM. Although discontinuities are apparent at some points along the
while still requiring zones of adhesion IM and OM which might suggest the novel membranes are protruding from these points,
this method does not support visualization at high enough resolution to determine where
between the inner and outer mem- LPS is accumulating definitively.
branes. For instance, Bayer's Bridges
might be periplasm-spanning proteins
that act as a scaffold for soluble periplasmic terial activity, and thus allow the postponement
shuttles. of efforts to deal with permeability problems
that those candidate antibiotics might have.
Imp is a large OMP with -barrel topology.
LPS Assembly in the OM
Using a strain in which Imp could be depleted,
We recently identified two factors that are in- we found that the protein is essential for E. coli
volved in assembling LPS at the cell surface-- viability. To study Imp further, we radiolabeled
namely, the -barrel protein Imp and the li- proteins and lipids after the mutants were de-
poprotein RlpB. The imp gene was first pleted of Imp, and then we collected membrane
identified in a screen for genes that allow E. coli fractions by sucrose density gradient centrifuga-
strains lacking maltoporin (the lamB gene prod- tion. Almost all proteins and lipids that are
uct) to use maltodextrins as a carbon source. made when Imp is depleted accumulate in a
The first imp mutants were in-frame deletions fraction that is denser than is the wild-type OM.
that disrupt the barrier function of the OM This result implicates Imp in OM biogenesis
(hence, the term imp, for increased membrane without, however, revealing its function.
permeability). These mutants are used in the LPS is essential for almost all gram-negative
pharmaceutical industry to test the intracellular bacteria, but not Neisseria meningitidis. Simi-
effectiveness of novel compounds with antibac- larly, imp is not essential in this bacterium.
Volume 2, Number 6, 2007 / Microbe Y 293
Taking advantage of this fact, Tommassen's together or independently. In the one-step sce-
group created a mutant strain that is lacking nario, LPS is inserted directly into the outer
imp. This strain has very low levels of LPS, and leaflet. Here, Imp and RlpB would act together
what little LPS remains in the cell does not reach to receive LPS at the inner face of the OM,
the cell surface and is unavailable to an enzyme transport it across the OM, and release it into
that is active on LPS in the OM. the outer leaflet. If depleting either Imp or RlpB
Using an affinity-tagged version of Imp, we blocks LPS delivery to the OM (i.e., LPS is never
showed that it interacts with the lipoprotein inserted into the inner leaflet if the two-step
RlpB in E. coli. Reciprocal purifications using hypothesis is correct and never inserted into the
affinity-tagged RlpB confirm this interaction, outer leaflet in the one-step model), we would
suggesting that Imp and RlpB interact in vivo. expect LPS to accumulate in the IM because it is
Although RlpB was previously known only as energetically unfavorable for LPS to exist as a
a low-abundance lipoprotein in E. coli (RlpB soluble entity in the periplasm. Alternatively, if
stands for rare lipoprotein B), we proved that it depleting Imp or RlpB abrogates the hypotheti-
is essential for viability and that it localizes to cal flipping function, LPS would accumulate in
the OM. To show that newly synthesized LPS the inner leaflet of the OM (Fig. 3).
never reaches the OM in Imp- or RlpB-depleted In Neisseria meningitidis, Imp null mutants
cells, we used PagP, an enzyme that modifies have severely reduced LPS levels. However, the
LPS in the OM by converting the hexa-acyl form opposite is true in E. coli. Tracking immuno-
of lipid A to the hepta-acyl form. blots with an antibody that recognizes LPS, LPS
In cells depleted of Imp or RlpB, PagP can levels increase as Imp and RlpB decrease. Con-
modify LPS that is synthesized before depletion, sistent with this observation, electron micro-
indicating that LPS is present in the outer leaflet graphs reveal that membrane material accumu-
of the OM. In contrast, PagP does not modify lates in such cells (Fig. 4). These pictures look
LPS synthesized after Imp or RlpB depletion, strikingly similar to those of ftsH mutants in
indicating that those LPS molecules never which LPS levels increase.
reached the outer leaflet. Because Imp targeting In our electron micrographs of Imp- and
and folding appear unaltered in RlpB-depleted RlpB-depleted cells, the extra membranes that
cells, the defects observed appear to be due to we observe appear in the periplasm (Fig. 4).
the role RlpB plays in LPS biogenesis. However, they could not be visualized at high-
enough resolution to determine whether they
The Mechanism of Imp/RlpB Action originated from the IM or OM. Accordingly, we
We think that LPS reaches the outer leaflet of the do not know whether LPS is accumulating in the
OM either by a one- or two-step process. In the IM or OM in these strains. Thus, while we know
two-step hypothesis, LPS is inserted into the that the Imp/RlpB complex is required to assem-
inner leaflet first, and then flipped to the outer ble LPS at the cell surface, we still need to
leaflet. Imp and RlpB could each be the OM determine whether one or both of these proteins
docking site, the inner-leaflet insertase, or the also functions as a delivery site for LPS in the
flippase, and they could perform these functions OM.
SUGGESTED READING
Bayer, M. E. 1968. Areas of adhesion between wall and membrane of Escherichia coli. J. Gen. Microbiol. 53:395 404.
Braun, M., and T. J. Silhavy. 2002. Imp/OstA is required for cell envelope biogenesis in Escherichia coli. Mol. Microbiol.
45:1289 1302.
Narita, S., S. Matsuyama, and H. Tokuda. 2004. Lipoprotein trafficking in Escherichia coli. Arch. Microbiol. 182:1 6.
Nikaido, H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol. Mol. Biol. Rev. 67:593
656.
Ruiz, N., D. Kahne, and T. J. Silhavy. 2006. Advances in understanding bacterial outer-membrane biogenesis. Nature Rev.
Microbiol. 4:57 66.
Raetz, C. R., and C. Whitfield. 2002. Lipopolysaccharide biogenesis. Annu. Rev. Biochem. 71:635700.
Sperandeo, P., R. Cescutti, R. Villa, C. Di Benedetto, D. Candia, G. Deho, and A. Polissi. 2006. Characterization of lptA and
lptB, two essential genes implicated in lipopolysaccharide transport to the outer membrane of Escherichia coli. J. Bacteriol.
189:244 253.
294 Y Microbe / Volume 2, Number 6, 2007
Tefsen, B., J. Geurtsen, F. Beckers, J. Tommassen, and H. de Cock. 2005. Lipopolysaccharide transport to the bacterial outer
membrane in spheroplasts. J. Biol. Chem. 280: 4504 4509.
Wu, T., A. C. McCandlish, L. S. Groenberger, C. C. Chng, T. J. Silhavy, and D. Kahne. 2006. Identification of a protein
complex that assembles lipopolysaccharide in the outer membrane of Escherichia coli. Proc. Natl. Acad. Sci. USA 103:11754-
11759.
Q: When does
4 x $17 = $48?
A: When Professors,
Postdoctoral Mentors and
Membership University Department
Recognizing the unique status of
scientists during the postdoctoral
Chairs purchase 4 or more
training phase of their careers, ASM
now offers Postdoctoral Membership.
student memberships for
Now postdoctoral trainees can receive all the
their
privileges* of Full Membership for $37 a year, for 4
consecutive years two years longer than the
Transitional Member rate offers.
undergraduate and
The 2007 membership fee will be $37 ($35 for graduate scholars!
those who join or renew online). Any microbiol-
ogist who has earned a doctorate within the past Buy 4 Student Memberships and receive a discount of $5
12 months is eligible. The Postdoctoral rate may per membership!
not be combined with the Transitional rate to
allow 6 years of reduced rate membership. To take advantage of this offer or to get more
information contact Membership Services at
Contact Membership Services at service@asmusa. membership@asmusa.org. You can also go to the
org for more information, or join online at Membership Menu to get started
estore.asm.org today! http://www.asm.org/Membership/index.asp?bid+363.
1752 N Street, NW *except the right to 1752 N Street, NW
Washington, DC 20036 vote or hold office
Washington, DC 20036
http://www.asm.org
service@asmusa.org
http://www.asm.org
membership@asmusa.org
Volume 2, Number 6, 2007 / Microbe Y 295
