Lipopolysaccharide Outer Core Is a Ligand for Corneal
Cell Binding and Ingestion of Pseudomonas aeruginosa
Tanweer S. Zaidi, Suzanne M. J. Fleiszig, Michael J. Preston, Joanna B. Goldberg,
and Gerald B. Pier
Purpose. Pseudomonas aeruginosa has been observed to be adherent to and inside epithelial
cells during experimental corneal infection. The authors identified bacterial ligands involved
in adherence and entry of P. aeruginosa into corneal epithelial cells.
Methods. In vitro gentamicin survival assays were used to determine the intracellular survival
of a panel of P. aeruginosa mutants. Strains (106 to 107 colony-forming units) were added to
primary cultures of rabbit corneal epithelial cells (~105/well) for 3 hours, nonadherent
bacteria were washed away, and extracellular bacteria were killed with gentamicin. The antibiotic was then washed away, and epithelial cells were lysed with 0.5% Triton X-100 to release
internalized bacteria. Bacterial association (sum of bound and internalized bacteria) was
measured by the omission of gentamicin. Similar assays were carried out with whole mouse
eyes in situ.
Results. A lipopolysaccharide core with an exposed terminal glucose residue was found to be
necessary for maximal association and entry of P. aeruginosa into corneal cells. Bacterial pili
and flagella were not involved. Mutants of P. aeruginosa strains that do not produce an LPS
core with a terminal glucose residue had a significantly lower level of association with (~50%)
and ingestion by (>90%, P < 0.01) corneal cells than did strains with this characteristic.
Complementation of the LPS production defect by plasmid-borne DNA returned association
and ingestion to near parental levels. Lipopolysaccharides and delipidated oligosaccharides
with a terminal glucose residue in the core inhibited bacterial association and entry into
corneal cells. Experiments using P. aeruginosa LPS mutants and corneal cells on whole mouse
eyes confirmed the role of the LPS core in cellular entry.
Conclusions. Corneal epithelial cells bind and internalize P. aeruginosa by the exposed LPS
core. Invest Ophthalmol Vis Sci. 1996; 37:976-986.
Aseudomonas aeruginosa is the pathogen most commonly involved in bacterial keratitis associated with
the use of contact lenses. 12 This form of corneal infection is rapidly progressive, is difficult to treat, and can
cause visual impairment. 1 ' 34 P. aeruginosa is generally
considered an extracellular pathogen, but we have
previously demonstrated that it is able to enter and
From the Charming Laboratory, Department of Medicine, Brigham and Women's
Hospital, Harvard Medical School, Boston, Massachusetts.
Presented in part as a poster at the annual meeting of the Association for Research
in Vision and Ophthalmology, Fort Lauderdale, Florida, May 1995, and at the
annual meeting of the American Society for Microbiology, Washington, DC, May
1995.
Supported by National Institutes of Health grants AI22535 (GBP), AI30050 and
AI35674 (JBG), EY06426 (MJP), and EY11221 (SMJF), and byaC.J. Martin
Fellowship from the Australian Government (SMJF).
Submitted for publication November 27, 1995; revised January 11, 1996; accepted
January 17, 1996.
Proprietary interest category: N.
Reprint requests: Gerald B. Pier, Channing Laboratory, Department of Medicine,
Brigham and Women's Hospital, 180 Longwood Avenue, Boston, MA 02115.
976
survive inside corneal cells in culture and in experimentally infected mice. 56 The entry into rabbit corneal cells during experimental contact lens-associated
infection7 and after corneal injury also has been reported. 8 These observations may explain why antibiotic therapy for P. aeruginosa keratitis is often ineffective. The ability to invade host cells is a virulence factor
of some bacterial species, including those in the genera of Salmonella, Shigella, Yersinia, and Neisseria.9'10
The ingestion by corneal epithelial cells of P. aeruginosa is related closely to the organism's ability to
attach to these cells.5 In addition, some strains of P.
aeruginosa, such as the commonly used 19660 strain,
are highly cytotoxic and readily kill corneal cells after
attachment and ingestion, rendering any ingested bacteria susceptible to gentamicin killing.5 Therefore, ingestion of cytotoxic P. aeruginosa strains is not readily
measured by the gentamicin-exclusion assay. Some reInvestigative Ophthalmology & Visual Science, May 1996, Vol. 37, No. 6
Copyright © Association for Research in Vision and Ophthalmology
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
Pseudomonas aeruginosa LPS and Cornea! Cell Ingestion
977
l. Strains of Pseudomonas aeruginosa and Plasmids
Used in This Study
TABLE
Strain or Plasmid Description
References
PAO1
PAOB1
AK1401
19,24
50
Wild-type strain, LPS-smooth, serogroup O5
Elastase (lasB)-negative mutant, lasB'-'Sl
Derivative of PAO1 LPS-defective, complete-core
plus 1 O side-chain repeat unit
AK1012
Derivative of PAO1 LPS-defective, incompletecore O side-chain deficient
algC mutant of PAO1 LPS-rough, incompletePAO1 algC' • tet
core O side-chain deficient
PAC1R
Wild-type strain, LPS-smooth, serogroup O3
PAC557
LPS-defective, complete-core derivative of
PAC1R O side-chain deficient
PAC611
Core-defective, O side-chain-expressing
derivative of PAC1R
PAC605
LPS-defective, incomplete-core derivative of
PAC1R O side-chain deficient
PAC1R algC'-: tet algC mutant of PAC1R, LPS-rough, incompletecore O side-chain deficient
6294
Corneal isolate, serogroup O6
6487
Corneal isolate, serogroup O6
PAK
PAK-fliC—
FAK-fliA—
PAK-NP
pLAFRl
pLPSl
pLPS188
Wild-type strain, LPS-smooth, serogroup O6
Nonmotile, mutation in flagellin gene,
fliC :Tn5G
Nonmotile, mutation in rpoF (JliA) gene,
fliA: :Tn5G
Nonadherent mutant, pilA:: tet
Tcr broad-host-range plasmid
algC gene cloned in pLAFRl
algC gene cloned in pUCP18
19,24
19,24
15
18
18
18
18
15
17
Bascom-Palmer Eye
Institute (Miami)
S. Lory, Seattle
44
51
27
52
20
20
LPS = lipopolysaccharide.
searchers 11 have proposed that lipopolysaccharide
(LPS), pili, 12 or both mediate the adherence of P.
aeruginosa to corneal tissue, binding specifically to gangliotetraosyl-ceramide (asialo GM1). 13 Others, 14 however, dispute whether asialo GM1 is present in the
corneas of rabbits and humans. Whether pili, LPS, or
asialo GM1 is involved in ingestion of P. aeruginosa by
corneal cells is unknown.
To identify the P. aeruginosa ligand(s) involved in
entry into corneal cells, we tested a variety of mutant
strains for their ability to invade primary cultures of
rabbit corneal epithelial cells. We found that the core
oligosaccharide fraction of the bacterial LPS was a
dominant bacterial ligand involved in corneal cell entry.
METHODS
Preparation of Bacteria
The bacterial strains and plasmids used in these experiments, along with their phenotypes and sources, are
listed in Table 1. Bacteria were grown overnight at
37°C on a trypticase soy agar plate. The inoculum was
prepared by resuspension of bacteria into Ham's F-12
medium to an appropriate optical density. The actual
colony-forming units (cfu) of bacteria added to a given
assay was determined by conventional methods of diluting and plating cells for enumeration.
Cell Cultures
Primary cultures of rabbit epithelial cells were prepared as previously described.5 In brief, rabbit eyes
were obtained from a commercial supplier (Pel Freeze
Biologicals, Rogers, AR) and washed in saline; the corneas were excised and then rinsed in Hanks' balanced
salt solution. The epithelium was removed after treatment with Dispase II (Boehringer Mannheim, Indianapolis, IN) in SHEM medium.5 All reagents, apart
from nutrient mixture Ham's F-12 (HyClone Laboratories, Logan, UT) and Eagle's minimal essential medium (Whittaker Bioproducts, Walkersville, MD),
were obtained from Sigma Chemical (St. Louis, MO).
Epithelial cells were seeded into tissue culture wells
(diameter, 15 mm; Costar, Cambridge, MA) in the
presence of SHEM medium and cultured at 37°C in
5% CO2. When epithelial cells became 95% to 100%
confluent, they were used in uptake experiments.
Association and Ingestion Assays
Cultured corneal epithelial cells (~105/well) were incubated with varying numbers of cfu of P. aeruginosa
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
978
Investigative Ophthalmology 8c Visual Science, May 1996, Vol. 37, No. 6
for 3 hours at 37°C in 5% CO2; subsequent washing
removed nonadherent bacteria. For quantifying (in
cfu) bacterial association with the cells, Triton X-100
(0.5%) was added to wells to lyse epithelial cells, and
the total number of P. aeruginosa organisms was determined by serial dilution and plating for bacterial enumeration. For measurement (in cfu) of the number
of P. aeruginosa organisms invading cells, gentamicin
(200 /zg/ml) was added to cultures after completion
of the association step to kill remaining extracellular
bacteria. After 2 hours, cells were washed thoroughly
to remove the antibiotic, and Triton X-100 (0.5%)
was added to lyse the cells and release intracellular P.
aeruginosa that survived antibiotic treatment. A viable
count of the surviving (ingested) bacteria released
from the cells was performed. In all assays, control
wells were included wherein Triton X-100 was added
to cells and was followed by 200 //g gentamicin/ml to
insure that this level of antibiotic could kill all the
bacteria. In addition, we determined that corneal cells
did not take up a sufficient amount of gentamicin to
kill 50 to 100 cfu of P. aeruginosa, which represented
<0.001% of the gentamicin added to the wells.
cells, debris was removed by centrifugation at 20,000g
for 15 minutes, and LPS was recovered from the supernatant by the addition of four volumes of 95% ethyl
alcohol. After 18 hours at 4°C, the precipitate was
collected by centrifugation and resuspended in PBS,
and this solution was ultracentrifuged for 3 hours ?t
100,000g. The LPS pellet obtained from the ultracentrifugation was resuspended in PBS; it was digested
with DNase and RNase (0.1 mg/ml) plus 4 //M CaCl2
and 4 //M MgCl2 overnight at 37°C and then with 0.1
mg of Pronase/ml for 2 hours at 56°C. The suspension
was ultracentrifuged again to pellet the LPS. This material was redissolved in water and lyophilized.
Core oligosaccharide fragments lacking lipid A
(delipidated LPS) were obtained by hydrolysis of LPS
in 1 % acetic acid for 3 hours at 95°C. The precipitated
lipid A was removed and the supernatant neutralized
for use in uptake assays. Inhibition of uptake by LPS
and oligosaccharides was measured by the addition of
the inhibitor to the ingestion assay mixture.
Statistical Analyses
Simple regression was used to analyze the interactions
of bacterial inoculum (log transformed) with associaIngestion by Epithelial Cells on Intact Mouse
tion and ingestion by epithelial cells. Analysis of variCorneas
ance (ANOVA) and the F-test were used to determine
Corneal surfaces of C57/B16 mouse eyes were injured
the significance level. Lines with slopes whose 99%
by scratching with a 26-gauge needle. All animals were
confidence intervals (CI) did not overlap were considtreated in accordance with the ARVO Statement for
ered significandy different. The 99% CI was used to
Use of Animals in Ophthalmic and Vision Research.
account for multigroup comparisons (i.e., application
The eyes were then enucleated and placed in a petri
of the Bonferroni correction to the significance
16
dish containing wet cotton. Each eye was inoculated
level
). Parametric statistics were used to analyze nor8
with 10 cfu of P. aeruginosa and was left at 37°C for 3 mally distributed data (either raw or log-transformed
hours. After this incubation period, the corneas were
results), including the unpaired and paired f-test for
washed extensively to remove nonadherent bacteria,
two-group comparisons and ANOVA for comparisons
and a viable count of the remaining adherent bacteria
of three or more groups. The Fisher PLSD statistic was
was performed after homogenization of some of the
used to determine which pairs of observations among
corneas in tryptic soy broth containing 0.5% Triton
groups of three or more observations were signifiX-100. This count gave us the number of bacteria asso- cantly different from each other at the 0.05 significiated with the cornea (total of adherent and ingested
cance level. For these analyses, the Statview SE+
bacteria).
Graphic software program (Abacus Concepts, BerkeIn an assessment of the level of ingestion of P.
ley, CA) was used on a Macintosh computer (Apple,
aeruginosa, additional corneas were washed and then Cupertino, CA).
treated with gentamicin for 1 hour to kill extracellular
bacteria. The antibiotic was washed away, the corneas
were homogenized in 0.5% Triton X-100 (as described
RESULTS
above), and a viable count was performed to deterEffect of Bacterial Inoculum on Adherence and
mine the number of internalized bacteria.
Ingestion by Epithelial Cells In Vitro
Preparation of Lipopolysaccharide
Before studying the ability of mutant strains of P. aeruLipopolysaccharide was obtained from P. aeruginosa ginosa derived from different parental backgrounds to
invade rabbit corneal epithelial cells, we determined
strains PAC557, which contains a complete LPS core
but no O-side chains, and PAC1R algC'-'-tet, a geneti- the effect of varying the bacterial inoculum on epithelial cell association and ingestion; the information encally derived strain that expresses an incomplete core
abled us to adjust the number of cfu of bacteria added
oligosaccharide and no O-side chains on the LPS.15
to cells to achieve comparable levels of ingestion with
After growth on agar plates for 24 hours, bacteria were
different strains and to compare laboratory strains
suspended in 10% sodium lauryl sarcosine to lyse the
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
Pseudomonas aeruginosa LPS and Cornea! Cell Ingestion
Log, p cfu
associated or
ingested
IO7'
Strain PAK
Association
10
979
Log, p cfu
associated or
ingested
,7,
Association
10
Slope=0.111
6
.
10
R 2 = 0.996
5
.
10
10
10
10
10
10" 10
10
Inoculum (cfu/ml)
10
Log, o cfu
associated or
ingested
Association
10
Slope = .053
R2 = 0.991
•
io5-
IO 3
IO 4
IO 5 IO 6 IO 7 IO 8
Inoculum (cfu/ml)
Log, 0 cfu
associated or
ingested
IO 9
Strain PAO1
Association
Slope = .105
R2 = 0.997
10
^
4.
• Ingestion
Slope = .004
R2 = 0.987
io ;
2
io :
10*
103
10°
10
Inoculum (cfu/ml)
Log, o cfu
associated or
ingested
7 ,
Ingestion
Slope = .00005
R2 = 0.942
2 .
IO7'
•
3
10
3.
a
io4-
IO110
4.
10
10
Strain 6294
107-|
Strain PAC1R
Ingestion
Slope = .007
R2 = 0.985
10
10'
Strain 6487
Association
Ingestion
Slope = .012
R2 = 0.997
103
10"
10
Inoculum (cfu/ml)
10
(i.e., PAO1, PAC1R, and PAK) with low-passage clinical isolates from corneal infection (i.e., 6294 and
6487). Results are presented in Figure 1. When the
log10 of the inoculum was plotted against the log10 of
bacteria associated with and ingested by corneal cells,
a dose-dependent increase in both parameters was
noted for all five strains, with correlation coefficients
(i?2 values) 2s0.942. For all five strains, the slopes of
the lines for association were steeper than the slopes
of the lines for ingestion, and the two slopes differed
significantly from each other (nonoverlapping 99%
CI). In general, the number of cfu invading the cells
was approximately 5% to 10% of the number associated with corneal cells, regardless of the inoculum
used. The exception involved strain PAC1R, for which
the figure was generally <1%. This strain had the
second highest slope for association (0.111) but a very
10
10
103
10°
10
Inoculum (cfu/ml)
10'
FIGURE 1. Effect of bacterial inoculum on
Pseudomonas aeruginosa association with
(upper lines) and ingestion by (lower lines)
rabbit corneal epithelial cells in primary
cultures. Each point represents the mean
of three or four determinations, and error bars show the standard deviations. All
slopes for ingestion and association had
nonoverlapping 99% confidence intervals, indicating that they differ at a level
of P < 0.01. Note that the range of the
x-axis is different for strains PAK and
PAC1R than for the other three strains.
low slope for ingestion (0.0005), indicating strong association with but poor ingestion by corneal cells.
In all cases, it was clear that there was a direct
relationship between the level of bacterial association
with the cells and subsequent ingestion by the cells.
Strains PAK and PAC1R were ingested poorly by corneal cells, with the lowest slopes for this measurement.
These strains required a > 10-fold higher inoculum to
achieve the same level of uptake as occurred with the
other three strains (Fig. 1). Strains PAO1, 6294, and
6487 differed significantly from each other in regard
to epithelial cell entry, with nonoverlapping 99% CI
for the slopes of the dose-response relationship of inoculum to ingestion. It is interesting that strains PAO1
and 6294 are highly infective in the murine cornealscratch model of ulcerative keratitis,17 whereas strains
PAK and PAC1R are not pathogenic in this model
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
980
Investigative Ophthalmology & Visual Science, May 1996, Vol. 37, No. 6
(unpublished observation, 1994). Strain 6487 has not
been evaluated in this model.
Bacterial Virulence Factors Involved in Uptake
In Vitro
We initially determined the role of pili, flagella, and elastase in the uptake by corneal cells of P. aeruginosa in vitro.
Strains lacking these factors (see Table 1) were ingested,
as were the isogenic parental strains expressing these factors (data not shown). In a comparison of the motile
flagellated strain with its nonmotile counterpart, the bacteria were spun gently (500g 10 minutes) onto the epithelial cell monolayers to negate the difference in the ability
of the two strains to swim toward targets. In contrast,
initial comparisons of strains expressing a smooth LPS
with isogenic derivatives expressing only a rough, partialcore LPS indicated that the strains expressing the rough
LPS were reduced severely in association with and ingestion by corneal cells (data not shown), which is comparable to results reported by Fletcher et al11 for adherence
of LPSsmooth and rough strains to intact and traumatized rat corneas.
Identification of the Iipopolysaccharide
Structure Needed for Corneal Cell Ingestion
The structure of the LPS cores of P. aeruginosa strains
PAC1R and PAO1 have been determined (Fig. 2), and
isogenic strains differing in the production of various
parts of the LPS core have been derived,18'19 which
would allow us to define precisely the LPS structure
promoting association and invasion with corneal cells.
In addition, we had available genetic constructs of
strains PAC1R and PAO1 wherein the algC gene,
which is needed for the production of a complete LPS
core,20 was inactivated by the insertion of a cassette
encoding tetracycline (tet) resistance into the coding
portion of this gene (strains PAC1R a/gC"tet and
PAO1 algC- '• tet15). Assessment of these strains for their
ability to be ingested by rabbit corneal cells in vitro
indicated that a terminal glucose residue linked to Nacetyl galactosamine had to be present on the LPS
core for efficient epithelial cell ingestion (Fig. 3).
Strain PAC557, which produces little or no serogroupspecific O-polysaccharide antigen,21 was as invasive as
both its parental strain and isogenic strain PAC611
(this latter strain also makes O-side chains, but with a
truncated core).21 Strain PAC557 similarly produces
little or no D-rhamnan common polysaccharide (also
known as "A-band" polysaccharide21), an additional
substituent on the P. aeruginosa LPS22; thus, this structure apparently was not involved in entry into corneal
cells.
Restoration of Epithelial Cell Association and
Uptake by Complementation of a Defective
Chromosomal algC Gene With an Intact,
Plasmid-borae Gene
P. aeruginosa strain AK1012 was derived by use of an
O-side chain-specific phage to select for a LPS-rough
mutant of P. aeruginosa PAO1.23'24 One mutation identified in this strain is in the algC gene, which encodes
an enzyme with both phosphomannomutase and phosphoglucomutase activity.15'25 Production of a smooth
or complete LPS can be restored in strain AK1012 by
a plasmid-borne copy of the algC gene.20 To confirm
that the disruption in algC function in P. aeruginosa
AK1012 was the only alteration leading to loss of the
capacity to invade corneal cells, we evaluated the ability of two plasmids (either pLPSl or pLPS188) with
intact copies of the algC gene to complement the genetic defect in strain AK1012 and to restore epithelial
cell invasiveness. As shown in Table 2, AK1012 organisms, carrying either complementing plasmid, associated with and invaded rabbit corneal epithelial cells
at a level comparable to that documented for the parental PAO1 strain. Both strain AK1012, carrying only
the cloning vector [AK1012 (pLAFRl)] and the genetically constructed strain PAO1 algC'- • tet, were reduced
significantly in their association and ingestion by epithelial cells.
Inhibition of Association and Ingestion Using
Iipopolysaccharide and the Delipidated Core
Oligosaccharide
The above results indicated that the exposed core oligosaccharide on LPS was the bacterial ligand used for
corneal cell association and ingestion. To evaluate this
hypothesis further, we tested the ability of intact LPS
and delipidated core oligosaccharide from P. aeruginosa PAC557 to inhibit the ingestion by corneal cells
of P. aerugmosa PAO1 in vitro. As shown in Figure 4,
the addition of 3=10 fig of intact LPS/ml or 2*1 fig of
delipidated LPS per milliliter to bacterial suspensions
during the ingestion assay significantly inhibited uptake of P. aeruginosa by corneal cells (P ^ 0.028, ANOVA, repeated measures). Doses of 50 to 100 fig of
intact or delipidated LPS of P. aeruginosa strain
PAC557 per milliliter were optimal for this purpose.
Intact and delipidated LPS were next evaluated
for their ability to inhibit corneal cell association and
ingestion of wild-type, LPS-smooth strains of P. aeruginosa and of strains expressing an intact LPS core but
devoid of O-side-chain expression. For a control, we
used LPS isolated from P. aeruginosa strain PAC1R
algC''-tet, which makes an incomplete LPS core. As
shown in Figure 5, the addition of 100 fig of intact or
delipidated core LPS from P. aeruginosa PAC557 per
milliliter to bacterial suspensions during the association and ingestion assays significantly inhibited corneal cell association and uptake of P. aeruginosa strains
(P < 0.01, ANOVA). In contrast, LPS and delipidated
LPS from P. aeruginosa PAC1R algC''-tet (incompletecore LPS) failed to inhibit significantly either association or uptake in vitro. To confirm that it is the core
oligosaccharide portion of the LPS that promotes uptake of P. aeruginosa, we tried to inhibit ingestion with
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
Pseudomonas aeruginosa LPS and Cornea! Cell Ingestion
981
D-GlC/7
1
D-Glcp
L-Rha
k
PO4
I—PAC611
I2
t
L
i
P
\
ot-D-Glc/?-( l-j->3 )-a-D-Glcp-(l-4>4)-a-D-GalpN-( 1—>3)-L-a-D-Hepp-( 1—>3 or 4)-L-a-D-Hepp-(1— >5)-<x—KD0/?-(2—> d
I
PAC557
AK1012
'
PAC605
PAOl algC::tet
PACIR a/gC: :tet
O-antigen with
1 repeat unit
AK1401
I
a
A
PO4
2
KDOp
Alanine
;
Polymerized
O-side chain
FIGURE 2. Structure of the lipopolysaccharide (LPS) core of Pseudomonas aeruginosa strains
PACIR and PAOl and depiction of the LPS structures for the mutant strains used in this
study. The structure represents a compilation of data from five sources.181946"49 Dashed,
diagonal arrows show a proposed alternative linkage between the upward projecting side
chains and the oligosaccharide backbone.18 Structural variation has been shown to exist
among the core oligosaccharides of different serogroups; the actual structures of the mutants
used probably vary from those shown in some of the linkages between adjacent sugars.
purified, high molecular weight O-polysaccharide side
chains from P. aeruginosa strains PACIR and PAOl.
Neither of these preparations inhibited bacterial ingestion (data not shown).
Role of Lipopolysaccharide Core
Oligosaccharide in Adherence and Ingestion by
Corneal Cells on Intact Mouse Eyes
To correlate our in vitro findings with the pathophysiology in an intact eye, we studied association and ingestion by epithelial cells on the scratch-injured corneas of enucleated whole mouse eyes of P. aeruginosa
strains expressing smooth, complete-core, or incomplete-core LPS. We used enucleated eyes because
strains expressing LPS-core oligosaccharides in the absence of O antigens are killed rapidly by complement
factors shortly after inoculation into the eye of a live
animal.17 Using the enucleated intact eye, we confirmed that P. aeruginosa strains expressing a completecore oligosaccharide adhered well to corneal tissues
and were ingested by epithelial cells at a level comparable to that documented for LPS-smooth parental
strains (Fig. 6). As in cell cultures, the strains expressing an incomplete LPS core had significantly reduced
ingestion by epithelial cells on scratch-injured corneas
of whole mouse eyes (P *£ 0.02, ANOVA). Interestingly, in these experiments, the association with intact
corneal tissues of strain PAC605 expressing an incomplete LPS core was reduced, but not significantly, from
the parental level. In addition, strain AK1401, expressing a complete-core oligosaccharide and a single Oantigen repeat, had a comparable level of association
with the intact cornea, as did parental strain PAOl,
but it had a significantly reduced level of entry into
corneal cells compared with this strain. These results
likely reflect a variable effect of LPS-core on adherence and entry of P. aeruginosa into mouse corneal
cells because of multiple receptor—ligand interactions, differences between rabbit and mouse cells, and
differences between in situ and in vitro situations.
DISCUSSION
The initial interaction of a bacterial pathogen with
components on the mucosal surfaces of humans is a
key part of the pathogenic process. The outcome of
this interaction will determine whether infection leads
to disease or to clearance of the pathogen. Many
pathogens use an array of surface factors to attach
to host tissues, including pili, flagella, matrix-binding
proteins, nonpilus adhesins, and LPS oligosaccharides.26"30 The host counters with mucous secretions,
glycoprotein barriers, phagocytes, immunoglobulins,
and inhibitors of bacterial binding.31"33 In corneal infections caused by P. aeruginosa, pili and LPS have
been reported to mediate bacterial adherence to corneal cells,12'34 but there is controversy about the epithelial cell receptor involved in adherence.35'36 Studies
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
982
Investigative Ophthalmology & Visual Science, May 1996, Vol. 37, No. 6
Strain
Association
FIGURE 3. Pseudomonas
Ingestion
Strain
i—I1
PACIR
PAC1F WM
PAC557
PAC557
PAC611
PAC611 1 | | |
B |
in
PAC605
PAC605 WM
PACIR B Hi
PACIR
105
10°
10
Mean cfu associating with cells
Association
aleC: :tet pBi
10 4
10
103
Mean cfu ingested by cells
Ingestion
Strain
PAO1
AK1401
AK1401
AK1012
AK1012
PAO1
a/gC::tet
PAO1
algC "tet
10"*
10 J
10"
Mean cfu associating with cells
10"
10 J
10*
Mean cfu ingested by cells
in our laboratory5'6 during the past few years also have
suggested that in addition to epithelial cell binding,
uptake could be another important element in the
pathogenesis of P. aeruginosa corneal infections. To
assess the hypothesis that bacterial ingestion by corneal epithelial cells is an important feature of the
pathogenic process, we have been conducting a series
of investigations to define bacterial and host factors
involved in ingestion. In this report of our studies,
using in vitro primary cultures of rabbit corneal epithelial cells and enucleated whole mouse eyes, we
identify the outer core of the bacterial LPS as a ligand
promoting binding to and ingestion by corneal cells
of P. aeruginosa. This finding is consistent with the
reports of Fletcher et al11 and Gupta et al12 indicating
a role for P. aeruginosa LPS in adherence to corneal
tissues. Gupta et al12 also proposed that the corneal
cell receptor for P. aeruginosa LPS is asialo GMl, which
is thought to be a receptor for pathogenic bacteria
in many host tissues.13'37'38 In addition, Gupta et al12
reported that P. aeruginosa pili bind to asialo GMl;
however, we found no role for pili in the ingestion of
this organism by cultured rabbit corneal epithelial
cells. We did note that bacterial uptake was much
more dependent on outer-core LPS than was association (i.e., adherent plus ingested bacteria); thus, ad-
10
aeru-
ginosa association with and
ingestion by rabbit corneal
epithelial cells in primary
cultures. Bars represent
means and error bars the
standard errors. Association
and ingestion of strains with
incomplete lipopolysaccharide (LPS) cores (i.e.,
PAC605, PACIR algC:: tet,
AK1012, and PAO1 algC-tet) was significantly less
than of strains with complete LPS cores (P< 0.001,
analyis of variance; P <
0.01, Fisher PLSD for all
pairwise comparisons between strains with incomplete LPS cores and all
strains expressing a complete core oligosaccharide).
In addition, strain PACIR
algC- '• tet was ingested significandy less than strain
PAC605 (P < 0.01 Fisher
PLSD). The number of colony-forming units added
per well was ~10 7 for the
PACIR series and ~10 6 for
the PAO1 series.
herence probably involves bacterial ligands in addition to LPS.
Because the interaction of a bacterium with an
epithelial cell is a complex process involving multiple
receptor-ligand interactions, we first characterized
the ability of five strains of P. aeruginosa to adhere to
and be ingested by cultured rabbit corneal epithelial
cells. We found substantial differences among the five
strains in the slopes of the lines relating increasing
bacterial inoculum to epithelial cell association and
ingestion. In general, however, association and ingestion were parallel phenomena: Those strains with
larger slopes for association also had larger slopes for
ingestion; the exception was strain PACIR, which had
a slope for association approximately 10-fold higher
than that of strain PAK, but it had a 10-fold lower
slope for ingestion. This information enabled us to
adjust the inocula of the different bacterial strains
so that their levels of association and ingestion were
comparable in studies with cultured rabbit corneal
epithelial cells.
In spite of strain-related variability, the experiments reported here consistently showed that the
outer-core oligosaccharide of P. aeruginosa LPS is one
of the ligands promoting bacterial association and ingestion in this system. In addition, the LPS outer-core
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
Pseudomonas aeruginosa LPS and Cornea! Cell Ingestion
983
TABLE 2.
Complementation of the Incomplete LPS Core Genetic
Defect in Pseudomonas aeruginosa Strain AK1012 Restores Epithelial
Cell Association and Ingestion
LPS
Strain*
Phenotype
(fu (± SEM) Associating
PAO1
PAOla/gC::tet
AK1012(pLAFRl)
AK1012(pLPSl)
AK1012(pLPS188)
Complete
Incomplete
Incomplete
Complete
Complete
2.8
6
1.1
2.4
3.0
X
X
X
X
X
105
103
105
105
105
±
±
±
±
±
3.1
1.6
4.0
4.1
3.5
X 104
X 10 3 ft
X 103f
X 104§
X 104§
tfu (± SEM) Invading
9.3
5.0
6.0
5.4
9.1
X
X
X
X
X
103 ±
102 ±
102 ±
103 ±
103 ±
1.8
1.4
1.1
2.7
1.6
X 103
X 102f
X 102f
X 102§
X 103§
* Approximately 10fi cfu of bacteria were added to each well.
t Strains PAO1 algC-'tet and AK1012(pLAFRl) are significantly less well associated with and ingested
by corneal epithelial cells than are strains with a complete LPS (strains PAO1, AK1012[pLPSl] and
AK1012[pLPS188], P < 0.001, ANOVA; P < 0.01, Fisher PLSD for all pairwise comparisons).
X Strain PAO1 algC- '• tet had a significantly lower epithelial cell association, but not ingestion, than
did strain AK1012(pLAFRl) (P < 0.01, Fisher PLSD).
§The two complemented strains of AK1012 are not significantly different in their association and
ingestion properties from the parental strain PAO1.
LPS = lipopolysaccharide; cfu = colony-forming unit.
oligosaccharide is needed for full ingestion of P. aeruginosa by cells on enucleated mouse corneas. Using
two series of isogenic strains of P. aeruginosa with wellcharacterized LPS-defective variants, 181924 as well as
genetically denned LPS-defective variants created by
the insertion of a tetracycline-resistance cartridge into
the aZgCgene,1520 we clearly showed that those strains
expressing a complete outer-core oligosaccharide associated with and were ingested by cultured epithelial
Inhibitor
cells most efficiently. Complementation of the LPScore synthesis defect in P. aeruginosa strain AK1012 by
the cloned algC gene restored association with and
ingestion to parental levels. Purified complete-core
LPS and complete-core oligosaccharide inhibited association with and ingestion by cultured corneal cells of
various P. aeruginosa strains. These findings parallel
those found in studies39 with respiratory epithelial cell
cultures and are consistent with a role for lipooligosaccharide in the ingestion by cultured epithelial cells of
Neisseria gonorrhoeae.40
Unlike most gram-negative bacteria, P. aeruginosa
produces an LPS with O-side chains attached to only
10% to 30% of the outer core residues—an arrangeLPS core oligosacccharide
ment that leaves the majority of the outer-core oligosaccharide exposed.41'42 This arrangement likely accounts for the ability of the LPS outer-core structure
of P. aeruginosa to serve as a ligand to epithelial cells.
The nature of the receptor on the epithelial cell is not
defined clearly; as noted above, controversy surrounds
the report that asialo GM1 is present on corneal epithelial cells and serves as a receptor for P. aeruginosa
LPS.35'36 In addition, the tetrasaccharide portion of
asialo GM1 could be present on the surface of a cell
either as part of this glycolipid or as an oligosaccharide
0
1
10
100
substituent on a glycoprotein. Thus, the exact nature
Concentration (|ig/ml) of inhibitor
of the epithelial cell target for P. aeruginosa LPS-mediFIGURE 4. Inhibition of ingestion of Pseudomonas aeruginosa ated adherence requires further study.
PAO1 by corneal cells in the presence of either intact comThe identification of the bacterial ligand promotplete-core lipopolysaccharide (LPS) or delipidated (95°C, 3
ing epithelial cell ingestion would normally lead to
hours, 1% acetic acid) complete-core oligosaccharide isoexperiments that evaluate the ability of isogenic bactelated from P. aeruginosa strain PAC557. Points represent
rial mutants that vary in their phenotypic expression
means, and error bars represent the standard errors. Doses
of 2= 10 ug of either intact or delipidated core LPS per millili- of the ligand to cause disease in an appropriate setting. However, this approach is not feasible in this
ter significantly inhibit ingestion (P = 0.028 and 0.025, resituation because P. aeruginosa strains expressing only
spectively, analysis of variance, repeated measures; P< 0.05,
an incomplete-core LPS are highly sensitive to the
Fisher PLSD for pairwise comparisons of doses of 10,50, and
100 [jig of either inhibitor per milliliter with no inhibitor). bactericidal effects of complement and, therefore, are
Complete core LPS
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
984
Investigative Ophthalmology 8c Visual Science, May 1996, Vol. 37, No. 6
Structure of inhibitor added
Strain:
Inhibitor:
PAC1R PACSS7 PAC1R PACSS7
Intact IPS
Dellpldated LPS
•
None
B
&
Complete core
Incomplete core
PAC1R
PAC557
Intact LPS
PAC1R
cfu Ingested
by epithelial
cells
PAC557
Delipldated LPS
cfu associating
with epithelial
cells
800000
these processes. Most bacterial interactions with mammalian cells are multifactorial, and multiple surface
proteins and other structures logically would be expected to participate in this interaction. Numerous
investigations26 have suggested that bacterial pili are
prominent ligands for binding epithelial cells,
whereas other studies29'4344 have indicated that nonpilus adhesins produced by P. aeruginosa also mediate
bacterial binding to both cells and mucins. In the
studies described here, epithelial cell ingestion of bacterial strains that lack a complete outer-core LPS generally was reduced by >90% from levels documented
for isogenic wild-type strains, although we cannot exclude the possibility that, by altering the structure of
cfu per
cornea
10
10
Strain:
Inhibitor:
PAO1 AK1401
Intact LPS
PAO1 AK1401
Dellpldated LPS
PAO1 AK1401
Intact LPS
PAO1
AK1401
Dellpldated LPS
Bacterial
strain
7,
PAC1R
PAC557
PAC605
6.
LPS phenotype
Smooth; wild-type
Complete core
Incomplete core
10 5.
FIGURE 5.
Inhibition of association with (left) and ingestion
4.
by (right) rabbit comeal cells of strains of Pseudomonas aerugi- 10
nosa expressing either smooth lipopolysaccharide (LPS;
3.
10
PAC1R and PAO1) or complete-core LPS (PAC557 and
AK1401): effect of addition to the assays of no inhibitor,
10'
intact LPS, or delipidated LPS. Bars represent the means,
and error bars represent the standard errors. The strain
used in each experiment and the type of inhibitor (intact
cfu per
or delipidated LPS) are shown on the *-axis of each graph.
cornea
Analysis of variance was used to compare the number of
107
colony-forming units (cfu) of P. aeruginosa associating with
or invading comeal cells under the three circumstances (P
6.
10
=S 0.03). Pairwise comparisons among the three groups
treated with inhibitors (whose structures are indicated in
5.
10
the legend) were all < 0.05 (Fisher's PLSD statistic). The
6
number of cfu of P. aeruginosa added varied from 10 to 2
10 4 .
X 107 cfu/well of epithelial cells, depending on the strain
and on whether association or ingestion was being mea3.
sured.
10
virtually avirulent in models of corneal infection.17
The same point applies to strains that express a complete-core LPS but lack O-side chains. In addition,
because LPS-core oligosaccharides are involved in adherence of P. aeruginosa to corneal cells and because
strategies that interrupt their production or activity
would affect epithelial cell adherence as well as ingestion, it is difficult to determine the specific contribution of ingestion to pathogenesis. Thus, alternative
strategies will be needed to illuminate the role of epithelial cell ingestion in the pathogenesis of P. aeruginosa corneal infection.
Although these results indicate that the LPS outer
core of P. aeruginosa is a ligand for this organism's
binding to and ingestion by corneal cells, it is likely
that other receptor-ligand interactions contribute to
Ingestion
Association
D
Q
•
Bacterial
strain
LPS phenotype
PA01
AK1401
AK1012
Smooth; wild-type
Complete core
Incomplete core
10'
Association
Ingestion
FIGURE 6. Association and ingestion by corneal epithelial
cells on scratch-injured, enucleated mouse eyes of isogenic
lipopolysaccharide (LPS)-derivatives of Pseudomonas aeruginosa strains of the PAC1R series (upper graph) and the PAO1
series (lower graph). For both series of LPS derivatives, ingestion was significantly (P =£ 0.04, ANOVA of log-transformed
results) less for the incomplete-core strains compared with
the LPS smooth strains. Only for the PAO1 series was association significantly (P < 0.002, ANOVA of log-transformed
results) less for the strain expressing an incomplete-core
LPS. Bars represent geometric means, and error bars represent the SEM. *Results are significandy different from the
wild-type strain; f results are significandy different from the
strain expressing a complete-core LPS. Corneas were inoculated with ~10 8 colony-forming units of each strain. ANOVA
= analysis of variance.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
Pseudomonas aeruginosa LPS and Cornea! Cell Ingestion
the LPS on these strains, we also altered the expression
or presentation of other prominent surface factors.
The addition of purified complete-core LPS or oligosaccharide generally inhibited ingestion by approximately 50%, but we would not expect these types of
experiments to achieve the same level of reduction in
binding and ingestion as was achieved by use of bacterial mutant strains. Of note is the recent report 40 that
N. gonorrhoea requires a complete oligosaccharide substituent on lipid A to invade epithelial cells fully, yet
the production of proteins conferring the opacity phenotype onto these bacteria is required for epithelial
cell ingestion.45 Taken together, these findings indicate that specific receptor-ligand interactions involving bacterial surface LPS carbohydrates and receptors
on epithelial cells contribute to the initial encounter
of a bacterial cell and a host mucosal surface. Further
studies are required to determine how best to evaluate
the importance of LPS outer core-mediated internalization of P. aeruginosa into epithelial cells in the
pathogenesis of corneal infections. Such studies must
account for the co-dependency of epithelial cell adherence with ingestion and the role of LPS O side
chains in conferring serum resistance onto P. aeruginosa.
985
9.
10.
11.
12.
13.
14.
15.
KeyWords
16.
bacterial pathogenesis, corneal ulceration, epithelial cell ingestion, lipopolysaccharide, Pseudomonas keratitis
17.
References
1. Alfonso E, Mandelbaum S, Fox MJ, Forster RK Ulcerative keratitis associated with contact lens wear. AmJ
Opthalmol. 1986; 101:429-433.
2. Stapleton F, DartJKG, Seal DV, Matheson M. Epidemiology of Pseudomonas aeruginosa keratitis in contact
lens wearers. Epidemiol Infect. 1995; 114:395-402.
3. Poggio EC, Glynn RJ, Schein OD, et al. The incidence
of ulcerative keratitis among users of daily-wear and
extended-wear soft contact lenses. N Engl J Med.
1989;321:779-783.
4. Schein OD, Glynn RJ, Poggio EC, Seddon JM, Kenyon
KR. The relative risk of ulcerative keratitis among users of daily-wear and . extended-wear soft contact
lenses: A case-control study. N Engl J Med. 1989; 321:
773-778.
5. Fleiszig SMJ, Zaidi TS, Pier GB. Pseudomonas aeruginosa
invasion of and multiplication within corneal epithelial cells in vitro. Infect Immun. 1995; 63:4072-4077.
6. Fleiszig SMJ, Zaidi TS, Fletcher EL, Preston MJ, Pier
GB. Pseudomonas aeruginosa invades corneal epithelial
cells during experimental infection. Infect Immun.
1994; 62:3485-3493.
7. Lawin-Brussel CA, Refojo MF, Leong FL, Hanninen
L, Kenyon KR. Time course of experimental Pseudomonas aeruginosa keratitis in contact lens overwear. Arch
Ophthalmol 1990; 108:1012-1019.
8. Stern GA, Lubniewski A, Allen C. The interaction between Pseudomonas aeruginosa and the corneal epithe-
18.
19.
20.
21.
22.
23.
24.
lium: An electron microscopic study. Arch Ophthalmol.
1985;103:1221-1225.
Falkow S, Isberg RR, Portnoy DA. The interaction of
bacteria with mammalian cells. Annu Rev Cell Biol.
1992; 8:333-363.
Leung KY, Finlay BB. Intracellular replication is essential for the virulence of Salmonella typhimurium. Proc
NatlAcadSd USA. 1991;88:11470-11474.
Fletcher EL, Fleiszig SMJ, Brennan NA. Lipopolysaccharide in adherence of Pseudomonas aeruginosa to the
cornea and contact lenses. Invest Ophthalmol Vis Sci.
1993;34:1930-1936.
Gupta SK, Berk RS, Masinick S, Hazlett LD. Pili and
lipopolysaccharide of Pseudomonas aeruginosa bind to
the glycolipid asialo GM1. Infect Immun. 1994; 62:
4572-4579.
Hazlett LD, Masinick S, Barrett R, Rosol K Evidence
for asialo GM1 as a corneal glycolipid receptor for
Pseudomonas aeruginosa adhesion. Infect Immun.
1993;61:5164-5173.
Zhao Z, Panjwani N. Pseudomonas aeruginosa infection
of the cornea and asialo GM(1). Infect Immun.
1995;63:353-355.
Coyne MJ, Russell KS, Coyle CL, Goldberg JB. The
Pseudomonas aeruginosa algC gene encodes phosphoglucomutase, required for the synthesis of a complete
lipopolysaccharide core. J Bacteriol. 1994; 176:35003507.
Abt K, Problems of repeated significance testing. Cont
Clin Trials. 1981; 1:377-381.
Preston MJ, Fleiszig SMJ, Zaidi TS, et al. Rapid and
sensitive method for evaluating Pseudomonas aeruginosa
virulence factors during corneal infections in mice.
Infect Immun. 1995;63:3497-3501.
Rowe PS, Meadow PM. Structure of the core oligosaccharide from the lipopolysaccharide of Pseudomonas
aeruginosa PAC1R and its defective mutants. EurJBiochem. 1983; 132:329-337.
Kropinski AM, Chan LC, Milazzo FH. The extraction
and analysis of lipopolysaccharides from Pseudomonas
aeruginosa strain PAO and three rough mutants. Can
fMicrobiol. 1979; 25:390-398.
Goldberg JB, Hatano K, Pier GB. Synthesis of lipopolysaccharide O side chains by Pseudomonas aeruginosa
strain PAO1 requires the enzyme phosphomannomutase. J Bacteriol. 1993; 175:1605-1611.
Dasgupta T, DeKievit TR, Masoud H, et al. Characterization of lipopolysaccharide-dencient mutants of
Pseudomonas aeruginosa derived from serotypes O3,
O5, and O6. Infect Immun. 1994; 62:809-817.
Hatano K, Goldberg JB, Pier GB. Pseudomonas aeruginosa lipopolysaccharide: Evidence that the O side
chains and common antigens are on the same molecule, f Bacteriol. 1993; 175:5117-5128.
Jarrell KF, Kropinski AM. Isolation and characterization of a bacteriophage specific for the lipopolysaccharide of rough derivatives of Pseudomonas aeruginosa
strain PAO. J Virol. 1981; 38:529-538.
Jarrell K, Kropinski AM. The chemical composition
of the lipopolysaccharide from Pseudomonas aeruginosa
strain PAO and a spontaneously derived rough mutant. Microbios. 1977; 19:103-116.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017
986
Investigative Ophthalmology & Visual Science, May 1996, Vol. 37, No. 6
25. Ye RW, Zielinski NA, Chakrabarty AM. Purification
and characterization of phosphomannomutase/phosphoglucomutase from Pseudomonas aeruginosa involved in biosynthesis of both alginate and lipopolysaccharide. / Bacteriol. 1994; 176:4851-4857.
26. Prince A. Mini-review—Adhesins and receptors of
Pseudomonas aeruginosa associated with infection of the
respiratory tract. Microb Pathog. 1992; 13:251-260.
27. Saiman L, Ishimoto K, Lory S, Prince A. The effect of
piliation and exoproduct expression on the adherence of Pseudomonas aeruginosa to respiratory epithelial monolayers. / Infect Dis. 1990; 161:541 -548.
28. PattiJM, Allen BL, McGavin MJ, Hook M. MSCRAMMmediated adherence of microorganisms to host tissues. Annu Rev Microbiol. 1994; 48:585-617.
29. Ramphal R, Koo L, Ishimoto KS, et al. Adhesion of
Pseudomonas aeruginosa pilin-deficient mutants to mucin. Infect Immun. 1991;59:1307-1311.
30. Porat N, Apicella MA, Blake MS. A lipooligosaccharide-binding site on HepG2 cells similar to the gonococcal opacity-associated surface protein Opa. Infect
Immun. 1995;63:2164-2172.
31. AlverdyJC, Spitz J, Hecht G, Ghandi S. Causes and
consequences of bacterial adherence to mucosal epithelia during critical illness. New Horiz. 1994; 2:264272.
32. McGhee JR, Mestecky J, Dertzbaugh MT, et al. The
mucosal immune system—From fundamental concepts to vaccine development. Vaccine. 1992; 10:7588.
33. Phillips AD, Shah AR, Walkersmith JA. Neutrophil response to mucosal infection. / Med Microbiol. 1992;
36:318-320.
34. Rudner XL, Zheng Z, Berk RS, Irvin RT, Hazlett LD.
Corneal epithelial glycoproteins exhibit Pseudomonas
aeruginosa pilus binding activity. Invest Ophthalmol Vis
Sri. 1992; 33:2185-2193.
35. Panjwani N. Is asialo GM(1) present in rabbit and
human corneal epithelium? Reply. Infect Immun.
1995;63:3746-3747. Letter to Editor.
36. Hazlett LD. Is asialo GM(1) present in rabbit and
human corneal epithelium? Infect Immun. 1995; 63:
3746. Letter to Editor.
37. Krivan HC, Roberts DD, Ginsburg V. Many pulmonary
pathogenic bacteria bind specifically to the carbohydrate sequence GalNAc P l-4Gal found in some glycolipids. Proc Natl Acad Sd USA. 1988;85:6157-6161.
38. Saiman L, Prince A. Pseudomonas aeruginosa pili bind
to asialo GM1 which is increased on the surface of
cystic fibrosis epithelial cells. / Clin Invest. 1993; 92:
1875-1880.
39. Pier GB, Grout M, Zaidi TS, et al. Role of mutant
CFTR in hypersusceptibility of cystic fibrosis patients
to lung infections. Srience. 1996:271:64-67.
40. Schwan ET, Robertson BD, Brade H, Vanputten JPM.
Gonococcal rfaF mutants express Rd(2) chemotype
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
LPS and do not enter epithelial host cells. Molec Microbiol. 1995; 15:267-275.
Darveau RP, Hancock REW. Procedure for isolation of
bacterial lipopolysaccharides from both smooth and
rough Pseudomonas aeruginosa and Salmonella typhimurium strains. /Bacteriol. 1983; 155:831-838.
Drewry DT, Symes KC, Gray GW, Wilkinson SG. Studies of polysaccharide fractions from the lipopolysaccharide of Pseudomonas aeruginosa N.C.T.C. 1999. Biochemf. 1975; 149:93-106.
Cervin MA, Simpson DA, Smith AL, Lory S. Differences in eukaryotic cell binding of Pseudomonas. Microb
Pathog. 1994; 17:291-299.
Simpson DA, Ramphal R, Lory S. Genetic analysis of
Pseudomonas aeruginosa adherence—Distinct genetic
loci control attachment to epithelial cells and mucins.
Infect Immun. 1992;60:3771-3779.
Weel JFL, Hopman CTP, Van Putten JPM. In situ expression and localization of Neisseria gonorrhoeae opacity proteins in infected epithelial cells: apparent role
of Opa proteins in cellular invasion. / Exp Med.
1991; 173:1395-1405.
Altman E, Sadovskaya I, Lam JS, Richards JS. Structural analysis of the Pseudomonas aeruginosa IATS O5
lipopolysaccharide core oligosaccharide. Abstracts of
the Fourth International Symposium on Pseudomonas: Biotechnology and Molecular Biology. Vancouver, BC, Canada. 1993; 73.
Masoud H, Altman E, Richards JC, Lam JS. General
strategy for structural analysis of the oligosaccharide
region of lipooligosaccharides: Structure of the oligosaccharide component of Pseudomonas aeruginosalATS
serotype O6 mutant R5 rough-type lipopolysaccharide. Biochemistry. 1994; 33:10568-10578.
Dekievit TR, Lam JS. Monoclonal antibodies that distinguish inner core, outer core, and lipid A regions of
Pseudomonas aeruginosa lipopolysaccharide. / Bacteriol.
1994; 176:7129-7139.
Masoud H, Sadovskaya I, de Kievit T, Altman E, Richards JC, LamJS. Structural elucidation of the lipopolysaccharide core region of the O-chain-deficient mutant strain A28 from Pseudomonas aeruginosa serotype
O6 (International Antigenic Typing Scheme), f Bacteriol. 1995; 177:6718-6726.
Toder DS, Gambello MJ, Iglewski BH. Pseudomonas
aeruginosa lasA: A second elastase under the transcriptional control of lasR. Molec Microbiol. 1991; 5:20032010.
Starnbach MN, Lory S. The fliA (rpoF) gene of Pseudomonas aeruginosa encodes an alternative sigma factor
required for flagellin synthesis. Molec Microbiol.
1992;6:459-469.
Friedman AM, Long SR, Brown SE, Buikema WJ, Ausubel FM. Construction of a broad host range cosmid
cloning vector and its use in the genetic analysis of
Rhizobium mutants. Gene. 1982; 18:289-296.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933418/ on 06/18/2017