Microbiology (2005), 151, 2829–2839
DOI 10.1099/mic.0.27984-0
The Pseudomonas fluorescens SBW25 wrinkly
spreader biofilm requires attachment factor,
cellulose fibre and LPS interactions to maintain
strength and integrity
Andrew J. Spiers1 and Paul B. Rainey1,2
1
Department of Plant Sciences, University of Oxford, South Parks Road, Oxford OX1 3RB, UK
Correspondence
Andrew J. Spiers
2
[email protected]
Received 22 February 2005
Revised
13 July 2005
Accepted 13 July 2005
School of Biological Sciences, University of Auckland, Private Bag 92019, Auckland, New
Zealand
The wrinkly spreader (WS) isolate of Pseudomonas fluorescens SBW25 forms a substantial
biofilm at the air–liquid interface. The biofilm is composed of an extracellular partially acetylated
cellulose-fibre matrix, and previous mutagenesis of WS with mini-Tn5 had identified both the
regulatory and cellulose-biosynthetic operons. One uncharacterized WS mutant, WS-5, still
expressed cellulose but produced very weak biofilms. In this work, the mini-Tn5 insertion site in
WS-5 has been identified as being immediately upstream of the tol-pal operon. Like Tol-Pal mutants
of other Gram-negative bacteria, WS-5 showed a ‘leaky-membrane’ phenotype, including the
serendipitous ability to utilize sucrose, increased uptake of the hydrophilic dye propidium iodide,
and the loss of lipopolysaccharide (LPS) expression. WS-5 cells were altered in relative
hydrophobicity, and showed poorer recruitment and maintenance in the biofilm than WS. The
WS-5 biofilm was also less sensitive to chemical interference during development. However,
growth rate, cellulose expression and attachment were not significantly different between WS and
WS-5. Finally, WS-5 biofilms could be partially complemented with WS-4, a biofilm- and
attachment-deficient mutant that expressed LPS, resulting in a mixed biofilm with significantly
increased strength. These findings show that a major component of the WS air–liquid biofilm
strength results from the interactions between LPS and the cellulose matrix of the biofilm – and that
in the WS biofilm, cellulose fibres, attachment factor and LPS are required for biofilm development,
strength and integrity.
INTRODUCTION
Many bacteria are capable of forming large assemblages on
plant and animal surfaces and tissues, on biological detritus,
sediments, soils and other geological structures, as well as
suspended flocs in water columns. Although often hard
to investigate in situ, significant biological properties are
attributed to these assemblages, including co-operative
behaviour, competitive advantage, and defence against predators, antibiotics and immune systems, physical disturbance, etc. (for recent reviews see Costerton et al., 1995;
Davey & O’Toole, 2000; Wimpenny et al., 2000; LappinScott & Bass, 2001; Sutherland, 2001a; Wilson, 2001;
Donlan, 2002; Dunne, 2002; Morris & Monier, 2003;
Hall-Stoodley et al., 2004). These assemblages range from
the largely random aggregation of bacteria growing on
surfaces, in semi-solid environments or in constrained
Abbreviations: A–L, air–liquid; CR, Congo red; DOC, deoxycholic acid;
EPS, exopolysaccharide; Hr, relative hydrophobicity; MDM, maximum
deformation mass; PI, propidium iodide.
0002-7984 G 2005 SGM
volumes, to complex structures incorporating substantial
amounts of extracellular matrix material (Sutherland,
2001b; Hall-Stoodley & Stoodley, 2002; Stoodley et al.,
2002; Ghigo, 2003). This latter type of assemblage represents
biofilms in stricto senso, and the presence of structural
matrix material provides biofilms with a cohesive physical
identity that may be lacking both in colonies and in slime.
It is clear that the physical resilience of biofilms is the
result of multiple interactions between matrix components
(often exopolysaccharides, EPS), bacterial surface appendages (fimbriae, flagella and aggregation factors) and coatings (lipopolysaccharide, LPS) and the surface colonized by
the bacteria (see references above and Dalton & March,
1998; Sutherland, 2001b; Donlan, 2002; Götz, 2002). In the
case of the biofilms produced by Salmonella typhimurium
and Salmonella enteritidis rdar mutants, and by the Pseudomonas fluorescens SBW25 wrinkly spreader, the expression
of a cellulose matrix and a fimbrial-like attachment factor are the primary components contributing to biofilm
strength and integrity (Römling & Rohde, 1999; Zogaj et al.,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
Printed in Great Britain
2829
A. J. Spiers and P. B. Rainey
2001; Solano et al., 2002; Spiers et al., 2002, 2003). In each
case, biofilms develop at the air–liquid (A–L) interface and
are substantially larger and more robust than the archetypical submerged biofilm produced by many other bacteria,
for example, Pseudomonas aeruginosa (see Wilson, 2001).
The wrinkly spreader (WS) is a niche-specialist genotype
that colonizes the A–L interface of liquid cultures, forming
an A–L biofilm, and grows poorly in the liquid column.
It arises by spontaneous mutation from the ancestral
(smooth; SM), non-biofilm-forming P. fluorescens SBW25
strain, in spatially structured microcosms, and shows significant negative frequency-dependent fitness advantage
over the ancestral strain (Rainey & Travisano, 1998). Its
selective advantage is attributable to cooperation among
individual WS cells: overproduction of attachment factors,
while costly to individual cells, results in the interests of
individuals aligning with those of the group and allows
colonization of the oxygen-replete A–L interface (Rainey &
Rainey, 2003).
In an investigation of the genes required for biofilm formation by P. fluorescens WS (using one particular WS
isolate, PR1200; Spiers et al., 2002), mini-Tn5 mutagenesis
identified two major loci – the wsp chemosensory operon
encoding the response regulator WspR, and the wss cellulose
biosynthesis operon, which includes genes involved in the
partial acetylation of the cellulose matrix (Spiers et al., 2002,
2003). WspR is required for the expression of cellulose
and a putative curli or thin aggregative fimbriae (Tafi)like attachment factor, both of which are required for
normal WS biofilm development and colony formation. In
addition, the cellulose acetylation-defective mutant WS-18
(WS wssF : : mini-Tn5) was found to produce weak biofilms. These findings suggest that the physical integrity of
the WS biofilm results from the interactions between
cellulose fibres, between fibres and attachment factor, and
between attachment factor and the walls of the microcosm
vial. This last interaction is required during the first phase of
biofilm development when bacteria attach in the meniscus
region of the liquid culture to the glass vials. Subsequent
growth out over the A–L interface results in the characteristic WS biofilm (Spiers et al., 2003).
One of the previously identified WS mini-Tn5 mutants, WS5, was found not to be associated with either of the wsp or
wss operons (Spiers et al., 2002). WS-5 cells form an unusual
colony morphology that is intermediate between that of
WS and the non-biofilm-forming SM strain that produces
smooth colonies on agar plates. Colonies of WS-5 bind
Congo red, indicating that it expresses cellulose, and this
strain is able to form weak biofilms when incubated
in liquid microcosms. In this work, we reveal the genomic
location of the transposon insertion responsible for the
observed defects in WS-5 and show how defects in LPS
expression affect A–L biofilm strength by significantly
altering the cellulose matrix–attachment factor–bacterial
cell interactions required for the normal development of
WS biofilms.
METHODS
Bacterial strains, plasmids, culture media and growth conditions. The Pseudomonas fluorescens strains used in this work are
derivatives of P. fluorescens SBW25 (Table 1) and were grown using
King’s B (KB) medium (King et al., 1954) or minimal medium containing 20 mM sucrose at 28 uC. A KB microcosm consisted of a
35 ml Universal glass vial containing 6 ml KB, and was incubated
with the lid held loosely in place with porous tape. P. fluorescens
motility was assessed using 0?16 KB/0?3 % (w/v) soft-agar plates.
Escherichia coli DH5a (Gibco-BRL) and S17-1-lpir (Simon et al.,
1983) were used for DNA manipulation and conjugation. E. coli
strains were grown using LB medium at 37 uC. The plasmids pBS+
(Stratagene) and pGEM7f (Promega) were used for subcloning and
sequencing, and the P. fluorescens SBW25 cosmid library in E. coli
S17-1-lpir was from Rainey (1999). p34S-Km3 was from Dennis &
Zylstra (1998). WspR and WspR19 were expressed in trans using
pVSP61-wspR12-VTcR (wild-type WspR) and pVSP61-wspR19-VTcR
(WspR R129C); pVSP61-VTcR was used as a negative control
(Goymer, 2002). Antibiotics were used at the following concentrations: ampicillin, 100 mg ml21; chloramphenicol, 20 mg ml21; kanamycin, 25 mg ml21; piperacillin, 150 mg ml21; and tetracycline,
12?5 mg ml21. Congo red (CR) [also known as Direct Red (DR) 28;
Table 1. P. fluorescens SBW25 strains used in this work
Strain
SM
JB01
AS24
WS and mutants
WS
WS-4
WS-5
WS-13
WS-18
WS tolA
2830
Genotype
Source
Wild-type SBW25
SM but NPTII promoter overexpressing wss
Non-chemotactic biofilm-forming strain evolved from a SBW25 cheA mutant
Rainey & Bailey (1996)
Spiers et al. (2002)
A. Spiers
Biofilm-forming strain (PR1200) evolved from SM
WS wspR : : mini-Tn5; isolated from a mini-transposon screen of WS
WS tol : : mini-Tn5; isolated from a mini-transposon screen of WS; mini-Tn5
is immediately upstream of ybgC, the first gene of the tol cluster
WS wssB : : mini-Tn5; isolated from a mini-transposon screen of WS
WS wssF : : mini-Tn5; isolated from a mini-transposon screen of WS
WS tolA : : KmR; kanamycin-resistance cassette inserted into the SphI site of tolA
Spiers et al. (2002)
Spiers et al. (2002)
Spiers et al. (2002)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
Spiers et al. (2002)
Spiers et al. (2002)
This work
Microbiology 151
Involvement of LPS in the wrinkly spreader biofilm
Sigma] was added to KB plates to a final concentration of 0?001 %
(w/v). Direct Blue (DB) 1 (Chicago Sky Blue; Aldrich), 14 (Trypan
Blue/Congo Blue; Aldrich), 15 (Sigma) and 53 (Evans Blue,
Aldrich), Direct Red (DR) 2 (benzopurpurin, Aldrich) and CR
were added to KB microcosms to a final concentration of 0?01 %
(w/v).
General molecular biology methods. Standard molecular bio-
logy techniques were used according to current protocols or manufacturer’s instructions. Optical densities were determined using
4 mm-pathway plastic cuvettes and a Spectronic 20 Genesys
(Spectronic Instruments) spectrophotometer. Conjugation and electroporation were used to transfer DNA between or into E. coli and
P. fluorescens. Ligation mixtures were dialysed against deionized
water for 30 min using 0?20 mm HA MF-membrane filters (Millipore) before electroporation. Plasmid and cosmid DNA was isolated
from E. coli using Qiagen mini-spin kits. DNA was analysed by
TBE-agarose gel electrophoresis and ethidium bromide-staining. The
following subclones of the cosmid pAS256 were made for sequencing purposes: pAS257-260 and 263 were HindIII fragments in
pBS+; pAS261 and 262 were SalI fragments in pBS+; pAS265, 266,
268 and 269 were partial HindIII deletions of pAS260; and pAS267,
271 and 272 were SphI fragments in pGEM7f (further mapping
data are available on request). Nucleotide sequence was obtained by
automated sequencing using standard vector and sequence-specific
primers.
Construction of the WS tolA mutant. This mutant was pro-
duced by the insertion of a kanamycin-resistance (KmR) cassette
into tolA. This was achieved by cloning KmR from p34S-Km3 into
the SphI site of pAS258, a pBS+ clone containing the three HindIII
fragments covering the ybgC-tolB region (Fig. 1), to give pAS298.
This plasmid was electroporated into WS, and KmR integrants were
isolated. These were grown in non-selective medium then plated
onto KB plus kanamycin. Individual colonies were tested for piperacillin sensitivity (bla from pAS298 confers resistance to piperacillin,
Pp) indicating vector excision, and an appropriate KmR PpS isolate
chosen as WS tolA.
LPS analysis. LPS samples were obtained from cultures using
EDTA extraction. Overnight KB cultures were first diluted to an
OD600 of ~0?5 to equalize cell numbers. Cells from 2 ml of culture
were resuspended in 100 ml deionized water; 400 ml 250 mM EDTA
(pH 8?0) was added and the suspension vortexed vigorously for 5 s.
The suspension was incubated at 37 uC for 30 min and vortexed
every 10 min. The supernatant was recovered for analysis after centrifugation at 10 000 g for 5 min. Aliquots were examined using
18 % deoxycholic acid polyacrylamide gel electrophoresis (DOCPAGE) and silver staining (Bio-Rad) according to Reuhs et al.
(1998).
The monoclonal antibody (mAb) BC12-CA4 (Meyer & Dewey, 2000)
was used to detect LPS by ELISA assay. This mAb recognizes an
unidentified antigen from Botrytis cinerea and binding is inhibited by
rhamnose (Rha), suggesting that the target is a Rha-glycosylated
protein. It was tested as a possible mAb against pseudomonads as
these bacteria contain substantial Rha polymers as part of the conserved LPS A-band O-polysaccharide component (Rocchetta et al.,
1999). KB-grown cells were resuspended in PBS (20 mM sodium
phosphate, 150 mM NaCl, pH 8?0), adjusted to an OD600 of ~0?5 to
equalize cell numbers, and used to produce a 101–103 dilution series in
PBS. Aliquots of 100 ml were adsorbed to a MaxiSorp ELISA Plate
(Nunc) overnight at 4 uC. After washing with PBS/0?05 % (v/v) Tween20, mAb was added and incubated for 1 h at 37 uC. Bound mAb was
detected using anti-mouse polyvalent immunoglobulin peroxidase conjugate (Sigma) and TMB substrate (Adgen) and absorbance
measured at 450 nm. mAb binding was tested as DA450 OD60021,
having adjusted for cell numbers. Measurements were performed in
duplicate.
Microscopy and FACS. Propidium iodide (PI) (Sigma) was used
Fig. 1. The mini-Tn5 transposon in WS-5 is located 24 bp
upstream of the start of ybgC, the first gene of the tol-pal
operon. The functions of YbgC, E and F are unknown, but
the TolA, R, Q and Pal proteins are involved in maintaining the
functional integrity of the inner and outer cell membranes. The
tol-pal operon is highly conserved amongst Gram-negative bacteria, and the P. fluorescens SBW25 gene sequences show
high levels of homology with the genes from E. coli. Downstream of tol-pal is a series of tRNA genes, followed by nadA
(L-aspartate oxidase subunit). The regions sequenced in this
work are shown in black. Gene positions were subsequently
determined by in silico analysis of this data and of the
Wellcome Trust Sanger Institute SBW25 genome project database. SBW25 tRNA sequences (small triangles) were obtained
but were not mapped within the contig. The kanamycinresistance cassette inserted in tolA at the SphI site, and the
position of the mini-Tn5 insertion in WS-5, are indicated.
http://mic.sgmjournals.org
to assess the leaky membrane phenotype of mutants after Gaspar
et al. (2000). PI (20 mM) was added to KB cultures and incubated at
28 uC in the dark for 30 min before examination with an Olympus
BX50 epifluorescence microscope. Calcofluor (Fluorescent Whitener
28, Sigma) was used to assess the presence of cellulose in biofilms or
colony material. Calcofluor (10 mM) was added to samples resuspended in KB, then incubated at 28 uC for 2 h before washing with
fresh KB and subsequent examination. A fluorescence-activated
cell sorter (FACS) was used to determine the percentage of cells
stained with PI. Overnight KB cultures were diluted to an OD600 of
0?100–0?150 in fresh KB and incubated with PI for 30 min. FLT-2
(orange) fluorescence and scattering was measured for 100 000
events, and the percentage above a threshold determined by preliminary comparison between WS and WS-5 was recorded for each
culture.
Biofilm and cellulose assays. Bacterial attachment to the glass of
KB microcosms in the meniscus region was determined quantitatively using crystal violet as previously described (Spiers et al., 2003)
and presented as the relative attachment with respect to the WS biofilm [A570 OD570WS21]. The absolute strength of KB-grown biofilms was determined by placing glass balls in the centre of each
biofilm until it broke, sank or was ripped from the sides of the
microcosm vial, to determine the maximum deformation mass
(MDM) (grams) (Spiers et al., 2003). In the case of the complementation and chemical interference experiments, biofilms were incubated on the bench at 20–22 uC to minimize physical disturbance;
as a result, MDM values were lower than those obtained at 28 uC.
In the complementation experiments, overnight cultures of strains
to be tested were diluted to an OD600 of 1?00. KB microcosms were
then inoculated with 100 ml aliquots of single strains or a mixture of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
2831
A. J. Spiers and P. B. Rainey
two strains (in a total volume of 100 ml) in such a manner that each
test used the same total number of cells. In the interference experiments, 120 ml 500 mM EDTA or water was added to mature KBgrown biofilms at the meniscus in order to avoid disruption of the
biofilm. The EDTA was allowed to mix by diffusion over a period of
2 h before MDM were determined. A quantitative measure of cellulose production using CR was made according to Spiers et al. (2003)
using material from KB microcosms, and data presented as DA490
OD60021, adjusted for cell numbers.
Hydrophobicity assay. Relative hydrophobicity (Hr) of strains was
tested using the MATH assay (after van der Mei & Busscher, 2001).
KB-grown cells were resuspended in 4 ml KB or 10 mM potassium
phosphate (pH 5?0) to an OD600 of ~0?5 and 1 ml hexadecane was
added. The samples were vortexed for 5 s and then allowed to stand
for 20 min before the OD600 of the aqueous phase was determined
(OD600i). The samples were revortexed for 60 s, allowed to stand
for 20 min and the OD600 remeasured (OD600f). The ratio OD600f
OD600i21 gave the relative hydrophobicity (Hr) of cells of each
strain (Hr determined using KB reflects real differences in hydrophobicity experienced by various strains in KB microcosms, whereas Hr
determined using potassium phosphate reflects an arbitrary difference useful for comparative purposes).
Recruitment and maintenance assays. Bacterial recruitment to
the surface was assessed by using standard cuvettes containing 2 ml
KB. This resulted in a liquid column of 30 mm, with the surface
20 mm above the region in which OD600 was determined. KB-grown
cells were resuspended in 2 ml fresh KB with a OD600 of 0?2–0?3.
OD600 measurements were taken every 10 min, with the cuvette
remaining in place throughout. After 1 h, samples were mixed, and
the OD600 was remeasured to determine bacterial growth, and drift
checked with a blank KB sample. The mean relative OD600 was calculated using the data from the 40–60 min time-points. Bacterial
maintenance within biofilms was assessed by sampling standard KB
microcosms. Aliquots of 1 ml were taken from immediately below
biofilms and the OD600 determined. Microcosms were then vortexed
and a 1 ml aliquot removed to determine the total OD600 (OD600t).
The ratio OD600 OD600t21 gave the proportion of cells in the liquid
column.
Statistical analyses. All data are presented as the mean±standard
error (SE) where appropriate. Assays were performed with five to
eight replicates unless otherwise stated. ANOVA and Student’s
t-tests were performed using JMP Statistical Discovery Software
(SAS) and P values provided where necessary.
RESULTS
Initial phenotypic characterization of WS-5
WS-5 was isolated from a mini-Tn5 mutagenesis of the WS
strain and is defective in expression of the wrinkled colony
morphology typical of wrinkly spreaders (Spiers et al.,
2002). As the first step in this work, we characterized the
phenotype of WS-5 on agar plates and in liquid microcosms. WS-5 colonies on both KB and LB agar were smoothlike and did not show the normal wrinkled colony
morphology of the WS. After 1 day, the colonies typically
were smaller and more waxy-looking than those produced
by the SM strain, and over a period of 2–3 days became
less SM-like but never fully WS. WS-5 colonies stained
orange with CR on agar plates, and WS-5 produced
very weak biofilms in KB microcosms. Examination of
2832
Calcofluor-stained colony and biofilm material by fluorescent microscopy confirmed that WS-5 expressed cellulose.
Identification and analysis of the mini-Tn5
insertion site in WS-5
In order to determine the location and genetic identity of
the mini-Tn5 insertion site in WS-5, we screened a P. fluorescens SBW25 cosmid library for clones that complemented
WS-5 in trans and restored the WS phenotype on KB agar
plates. One cosmid was isolated (pAS256) and restriction
analysis indicated that it contained a ~20 kb fragment,
which was then randomly subcloned and end-sequences
obtained. This sequence-sampling allowed identification of
two well-conserved gene clusters at either end of the cosmid
insert: the gsv glycine cleavage system (Okamura-Ikeda et al.,
1993), and tol-pal (also referred to as tol-oprL) (Sturgis,
2001). When located on the unfinished SBW25 genome, the
sequences identified a single contig covering the entire
cosmid insert region. Using a mini-Tn5-specific primer and
nested-PCR sequencing, the insertion site of mini-Tn5 in
WS-5 was determined immediately upstream of ybgC, the
first gene in the tol-pal cluster (i.e. WS-5 was WS tol : : miniTn5) (Fig. 1).
Tol-Pal proteins are involved in the normal interaction of
the inner and outer membranes and are found throughout the eubacteria (Sturgis, 2001; Lazzaroni et al., 1999;
Lloubes et al., 2001). Tol-Pal system mutants typically show
impaired control of membrane channels (resulting in
problems with uptake or export, leakage of proteins from
the cytoplasm, sensitivity to pH and osmotic stress) and the
disruption of outer-membrane or cell-surface components,
including a reduction in or loss of LPS expression (Gaspar
et al., 2000).
From the position of the mini-Tn5 insertion site in WS-5
(WS tol : : mini-Tn5), it was clear that expression of ybgCtolQRAB-pal-ybgF would be severely affected (the Tol-Pal
system per se includes tolQRAB-pal; per contra no function
has been reported for ybgC or ybgF: Lloubes et al., 2001). In
order to confirm that the disruption of a known tol gene was
sufficient to explain the WS-5 phenotype, we made a WS
tolA mutant in which tolA was disrupted and downstream
tolB-ybgF expression compromised (Fig. 1). WS tolA2
showed a WS-5-like colony morphology, expressed cellulose
and produced weak biofilms in KB microcosms, suggesting
that it was the disruption of the functionally known tol
genes, rather than the functionally uncharacterized ybgC
gene, that was responsible for the phenotype of WS-5. In
contrast to WS-5 (WS tol : : mini-Tn5), WS tolA in shaking
cultures produced more floccular material, suggesting that
the tolA mutation generated a more severe phenotype
(through the inactivation of TolA and loss of Tol-Pal
function) than the WS-5 mini-Tn5 insertion (in which
tolQRAB-pal-ybgF expression was reduced, allowing some
Tol-Pal function to be preserved). For this reason, the rest of
our work focused on the comparative analysis of WS-5
alone.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
Microbiology 151
Involvement of LPS in the wrinkly spreader biofilm
Confirmation of the leaky-membrane phenotype
In order to determine whether WS-5 showed the leakymembrane phenotype typical of Tol-Pal mutants, we
assessed membrane integrity using the fluorescent DNAbinding dye propidium iodide (PI). This hydrophilic dye
cannot pass through the bacterial membrane, and can only
bind DNA if the membrane has been damaged. Fluorescent
microscopy of exponential-phase WS-5 cells grown in KB
with PI showed that a significant number of cells stained
with the dye (and cells were misshapen), indicating that WS5 shows the expected Tol-Pal leaky-membrane phenotype.
In contrast, most WS cells did not stain with PI and showed
no evidence of misshapen cell morphologies. We used FACS
analysis to quantify the relative differences in PI uptake, and
found that WS-5 (WS tol : : mini-Tn5) staining was 3?35fold greater than the mean uptake for SM, WS, WS-4 (WS
wspR : : mini-Tn5) and WS-18 (WS wssF : : mini-Tn5) cells.
Some Tol-Pal system mutants are able to utilize small
molecular mass molecules as sole carbon sources that diffuse
across the damaged membrane, which otherwise could
not cross into the cytoplasm, where they are metabolized
(Llamas et al., 2003). In a test of this particular phenotype,
we found that WS-5 was able to grow on minimal agar
supplemented with sucrose, whereas neither the SM nor
WS strains could use the disaccharide as the sole carbon
source. These findings are all consistent with the leakymembrane phenotype expected from the mini-Tn5 insertion site in WS-5.
system mutants, we considered that the loss of LPS expression was most likely to have an impact on biofilm formation.
We therefore directly tested this hypothesis by examining whether LPS expression was reduced in WS-5, whether
WS-5 cells showed altered hydrophobicity, and whether
WS biofilm strength could be changed by chemical interference targeted at LPS–cellulose fibre–attachment factor
interactions.
Expression of LPS
LPS expression is strongly reduced in Tol-Pal mutants
(Gaspar et al., 2000). In order to determine whether WS-5
showed a similar reduction in LPS expression, we prepared
LPS EDTA-extracts from overnight KB cultures in which cell
densities had been first equalized. These extracts were electrophoresed using DOC-PA gels that were then silverstained to reveal the major LPS bands (Fig. 2). WS-5 (WS
tol : : mini-Tn5) expressed insignificant amounts of LPS
when compared with either WS, WS-4 (WS wspR : : miniTn5) or WS-18 (WS wssF : : mini-Tn5). LPS levels in WS and
WS-5 were also investigated using the mAb BC12-CA4.
Although the binding of mAb BC12-CA4 to P. fluorescens
was weak, ELISA assays clearly showed a significantly greater
(5?76) mAb binding to WS than WS-5 cells (P=0?0384)
(DA450 OD60021±SE: WS, 0?554±0?022; WS-5, 0?097±
0?011), further supporting our DOC-PAGE observations
that WS-5 does not express detectable amounts of LPS.
Relative hydrophobicity of WS strains
WS-5 is insensitive to WspR reactivation of the
WS phenotype
In order to determine how a disruption of the Tol-Pal
system might result in weak biofilm formation by WS-5,
we first examined whether the WS phenotype in WS-5 could
be recovered by WspR expressed in trans. Previous work
has identified WspR as a regulator of both cellulose and
attachment-factor expression (Spiers et al., 2002, 2003).
When expressed in trans in SM, both wild-type WspR
(WspR12) and the constitutively active mutant WspR19
produce WS-like colony morphologies (Goymer, 2002). We
determined the colony phenotypes of SM and WS-5 (WS
tol : : mini-Tn5) carrying pVSP61-VTcR, pVSP61-wspR12R
R
VTc and pVSP61-wspR19-VTc on both KB and LB agar
plates. The control plasmid pVSP61-VTcR did not alter
either SM or WS-5 colony morphologies, and both pVSP61wspR12-VTcR and pVSP61-wspR19-VTcR produced WS-like
colonies in SM. In contrast, neither pVSP61-wspR12-VTcR
nor pVSP61-wspR19-VTcR altered the colony morphology
of WS-5.
These findings indicate that WS-5 is not a mutant in which
the WS phenotype has been turned off, or in which the
signal that activates the WS phenotype has been interrupted.
It therefore seemed most likely that a third component
required for normal WS biofilm formation was no longer
available. Of all of the phenotypes associated with Tol-Pal
http://mic.sgmjournals.org
LPS expression is known to affect the surface charge and/or
relative hydrophobicity (Hr) of bacterial cells (Rocchetta
WS
4
5
18
Fig. 2. LPS expression in WS-5 is strongly reduced in comparison with the wrinkly spreader and other mutants. From left
to right: LPS samples from WS, WS-4 (WS wspR : : mini-Tn5),
WS-5 (WS tol : : mini-Tn5) and WS-18 (WS wssF : : mini-Tn5).
The major LPS bands are indicated by triangles. LPS samples
were prepared from overnight KB cultures adjusted to give the
same OD600 cell density. Samples were extracted using EDTA,
electrophoresed in an 18 % DOC-polyacrylamide gel and then
silver-stained to detect LPS.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
2833
A. J. Spiers and P. B. Rainey
et al., 1999). We used the microbial adhesion to hydrocarbon (MATH) assay to determine whether WS-5 cells
showed altered Hr compared to WS. When the assay was
performed in KB, WS, WS-5 (WS tol : : mini-Tn5) and WS18 (WS wssF : : mini-Tn5) showed significantly different
Hr from one another (P=0?0005) (Hr±SE: WS, 0?0574±
0?0113; WS-5, 0?0959±0?0100; WS-18, 0?0163±0?0050).
By determining Hr in KB, we have measured the real
difference in cell hydrophobicities in the same environment
in which the biofilms are produced. In contrast, when the
assay was performed using potassium phosphate buffer
(to determine an arbitrary difference in Hr), WS-5 (WS
tol : : mini-Tn5) and WS-18 (WS wssF : : mini-Tn5) Hr
values were similar (P=0?9800), but different from WS
(P=0?001) (Hr±SE: WS, 0?9097±0?0075; WS-5, 0?9938±
0?0054; WS-18, 0?9934±0?0147). These findings show that
in the context of the KB microcosm, the surface charge and/
or hydrophobicity of WS, WS-5 and WS-18 bacterial cells
differ, and changes in the chemical environment cause
changes in relative hydrophobicity.
Comparison between WS and WS-5 biofilms
In order to quantify the differences in the physical characteristics between the WS-5 biofilm and those produced by
WS and other mutants, we determined the relative attachment and maximum deformation mass (MDM, strength) of
3-day-old KB-grown biofilms. Although the ability of WS-5
(WS tol : : mini-Tn5) to attach to the surface of the glass
microcosms was not significantly different from that of WS
or WS-18 (WS wssF : : mini-Tn5) (P=0?0547) (Fig. 3a), the
absolute strength of the WS-5 biofilm was substantially less
than that of WS or WS-18 (P<0?0001) (Fig. 3b). In order to
determine whether the reduced strength of WS-5 biofilms
was due to a decreased rate of growth, we measured growth
in KB over 24 h for WS, WS-5 (WS tol : : mini-Tn5) and
WS-18 (WS wssF : : mini-Tn5) (DOD600 h21±SE: WS,
0?058±0?001; WS-5, 0?071±0?001; WS-18, 0?068±
0?001). The three growth rates were not significantly
different (P=0?4254), indicating that the reduced WS-5
biofilm strength could not be due to a decreased rate of
growth. Finally, we also measured the relative amounts
of cellulose produced by each strain using a CR-binding
assay. WS and WS-5 (WS tol : : mini-Tn5) bound
similar amounts of CR (P=0?0965) and slightly less
(~1?266) than WS-18 (WS wssF : : mini-Tn5) (P=0?0049)
(DA490 OD60021±SE: WS, 0?317±0?019; WS-5, 0?368±
0?019; WS-18, 0?464±0?033). CR binds components other
than cellulose, and in the case of WS and WS mutants, CR is
bound by cellulose and attachment factor (Spiers et al.,
2003). Having demonstrated that WS, WS-5 and WS-18
show the same degree of attachment and similar levels of
CR-binding, we therefore conclude that the three strains
express similar levels of cellulose. This finding indicates that
the reduced strength of WS-5 biofilms is not due to a
reduced level of cellulose expression.
Recruitment to the A–L interface
The differences in relative hydrophobicity (Hr) between WS,
WS-5 and WS-18 cells might affect A–L biofilm strength by
enhancing the recruitment of cells to the meniscus region
where initial attachment to the glass surface of microcosms
occurs, and by maintaining cells within the developing
biofilm during growth. We examined differences in recruitment by monitoring OD600 at the bottom of the liquid
column using standard spectrophotometer cuvettes. We
reasoned that bacterial cells may adhere to the surface of
the cuvette, or remain a homogeneous suspension of cells,
and thus show no change in OD600. Alternatively, the cells
may be recruited to the surface (by chemotaxis or random
motion) and maintained through attachment to the walls
or aggregation with other bacteria at the surface. First,
however, we determined that WS, WS-4 (WS wspR : : miniTn5), WS-5 (WS tol : : mini-Tn5) and WS-18 (WS
wssF : : mini-Tn5) cells were motile by direct microscopic
examination, and significantly different from the nonchemotactic AS24 (an evolved SM cheA) (P<0?0001)
(migration through soft agar, mm h21: WS, 0?681±
0?019; WS-5, 0?394±0?019; WS-18, 0?594±0?019; cf.
Fig. 3. Relative attachment and strength of
WS-5 A–L biofilms. (a) Relative attachment
for SM, WS, WS-4 (WS wspR : : mini-Tn5),
WS-5 (WS tol : : mini-Tn5), WS-13 (WS
wssB : : mini-Tn5)
and
WS-18
(WS
wssF : : mini-Tn5) (OD570 OD570WS”1); (b)
Maximum deformation mass (MDM) for WS,
WS-5, and WS-18 (SM, WS-4 and WS-13
do not produce biofilms). Microcosms were
incubated at 28 6C for 3 days before assay.
Means±SE are shown.
2834
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
Microbiology 151
Involvement of LPS in the wrinkly spreader biofilm
WS-4, which is not impeded by attachment factor, 0?836±
0?022; and AS24, 0?125±0?019). The relative OD600 at the
bottom of KB liquid columns was found to decrease for WS,
WS-5 and WS-18 as cells migrated towards the surface over a
period of 1 h (Fig. 4). WS cells showed a significantly greater
(~1?56) level of recruitment to the surface than either
WS-5 (WS tol : : mini-Tn5) or WS-18 (WS wssF : : miniTn5) (P=0?006) (relative OD600±SE: WS, 0?8877±0?0013;
WS-5, 0?9224±0?0066; WS-18, 0?9299±0?0011). In contrast, WS-4 (WS wspR : : mini-Tn5) cells, unable to express
cellulose or attachment factor, showed a substantially different behaviour in which cells initially adhered to the sides of
the cuvette before slowly migrating towards the surface after
20 min. We also measured the ability of WS, WS-5 and WS18 biofilms to maintain cells within the developing biofilm
during growth. After 3 days, the proportion of cells found in
the liquid column under the biofilm was significantly greater
(3?7–4?46) for WS-5 (WS tol : : mini-Tn5) and WS-18 (WS
wssF : : mini-Tn5) than for WS (P=0?0001) (though the
total OD600 achieved by WS, WS-5 and WS-18 was the same,
P=0?5022). From these findings it is clear that the weaker
WS-5 and WS-18 biofilms are unable to recruit and maintain cells within the biofilm as efficiently as the WS.
Chemical interference of interactions amongst
biofilm components
Previously, we have shown that the presence of the dye CR
resulted in a significant decrease in the strength of WS
Fig. 4. WS-5 and WS-18 are unable to recruit cells into the
developing biofilm as efficiently as WS. Recruitment to the surface through the liquid KB column, shown by a decrease in
OD600, is faster for WS ($) than for WS-5 (WS tol : : mini-Tn5)
(n) or WS-18 (WS wssF : : mini-Tn5) (#). WS-4 cells (WS
wspR : : mini-Tn5) (%) show a different behaviour, in which cell
densities initially increase at the bottom of the cuvette (these
cells may be attached to the surface of the cuvette or they
might remain in suspension), and then slowly decrease over
time. Recruitment assays were carried out at 20–22 6C, during
which 3–5 % growth-dependent increases in final OD600 were
seen. Mean±SE values for 40–60 min are shown on the right.
http://mic.sgmjournals.org
biofilms by interfering between the normal cellulose fibre
and/or attachment factor interactions (Spiers et al., 2003).
In order to further elucidate interactions between biofilm
components, we tested WS, WS-5 and WS-18 biofilm
strengths in the presence of Ca2+, Fe3+, EDTA and various
diazo dyes structurally related to CR. The metal anions are
expected to bind EPS/LPS and alter the normal cell-surface
charge distribution; EDTA chelates Mg2+, which is known
to have a major role in LPS charge-neutralization (Groisman
et al., 1997; Rocchetta et al., 1999). In contrast, the diazo
dyes bind cellulose differentially due to slight structural
variations (Kai & Mondal, 1997), and are expected to
interact similarly with the WS attachment factor.
In initial tests, we found that both Ca2+ and Fe3+ severely
affected growth in KB microcosms, whereas the addition
of more Mg2+ (KB contains ~6 mM Mg2+) had no effect
on MDM, and therefore these were not tested further.
However, at low levels of EDTA (2–10 mM), a significant
decrease in WS and WS-18 (WS wssF : : mini-Tn5) MDM
was observed, whereas WS-5 (WS tol : : mini-Tn5) MDM
was not significantly affected even at 10 mM EDTA (P=
0?1333) (Fig. 5a). (5 mM EDTA had no effect on maximum
growth rate, but 10 mM EDTA resulted in a 0?26 reduction
in growth rate.) EDTA might act to prevent irreversible
interactions that occur during biofilm development, or it
might act to destabilize reversible interactions that maintain
biofilm strength. To determine which of these possibilities
was more likely, we tested the MDM of mature WS biofilms
2 h after EDTA had been added to a final concentration
of 10 mM. There was no significant difference in MDM
between the EDTA-treated biofilms and water-treated
negative controls (P=0?6708), indicating that EDTA affects
the establishment of irreversible interactions that form
during biofilm development, rather than destabilizing reversible interactions (i.e. the constant association and disassociation of cellulose fibres, attachment factor and LPS) that
might occur in the mature biofilm.
We also tested CR and related diazo dyes (DR 2, DB 1, 14,
15 and 53) (Fig. 5b), having first determined that no dye
showed a toxic effect on growth at the concentration tested.
None of the dyes increased the relative MDM of WS-5 (WS
tol : : mini-Tn5) (P=0?9804). However, DB 1 and DB 53
differentiated WS-5, DB14 differentiated WS-18 (WS
wssF : : mini-Tn5), and DR 2 differentiated WS from the
other two strains. These findings confirmed our expectations that EDTA and the diazo dyes would differentiate
between WS, WS-5 and WS-18 biofilms through differential
interference of interactions between biofilm components.
This strongly suggests that biofilm strength is the result of
multiple interactions between biofilm components, and that
WS-5 biofilms lack some component found in both WS and
WS-18 biofilms.
Complementation of WS-5 biofilms with WS-4
If biofilm development and final strength are the result of
multiple cellulose fibre–attachment factor–LPS interactions,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
2835
A. J. Spiers and P. B. Rainey
Fig. 5. EDTA and diazo dyes interfere with normal interactions between biofilm components and result in altered biofilm
strengths. (a) The addition of EDTA to standard KB microcosms had a significant effect on the relative maximum deformation
mass (MDM) of both WS ($) and WS-18 (WS wssF : : mini-Tn5) (#) biofilms, but no significant effect on the very weak WS5 (WS tol : : mini-Tn5) (n) biofilm except at high concentrations of EDTA. (b) Diazo dyes significantly altered the relative MDM
of WS (dark bars) and WS-18 (white bars) biofilms, but had little impact on WS-5 (light grey bars) biofilms. Ctrl, no dye
added. CR is also known as DR 28. Microcosms were incubated at 28 6C for 3 days before assay. Mean±SE of relative MDM
shown for both assays. Note the log scale in (b).
2836
0.035
0.030
1.7
+
0.020
+
0.025
MDM (g)
we reasoned that the weak WS-5 biofilm may be complemented by a second strain capable of expressing LPS,
but which cannot express cellulose or attachment factor
(e.g. WS-4). In order to test this expectation, we determined the MDM of KB-grown WS-4 (WS wspR : : miniTn5)/WS-5 (WS tol : : mini-Tn5) mixed biofilms (Fig. 6).
Biofilms produced from WS-5 alone or 1 : 9 WS-4/WS-5
were significantly weaker than 1 : 4 WS-4/WS-5-mixed
biofilms (P=0?0465, 0?0486), but none of the mixed
biofilms reached the strength of WS biofilms. Nevertheless,
this result demonstrates that WS-4 can partially complement WS-5 biofilm strength. Furthermore, similar partial
complementation of JB01 (SM NPTII : : wss) biofilms with
WS-13 (WS wssB : : mini-Tn5) was also seen (JB01 is an
SM derivative overexpressing cellulose but not expressing
attachment factor, which produces a particularly weak
biofilm: Spiers et al., 2002). When JB01 was complemented
with WS-13 expressing attachment factor, the MDM of
the 1 : 1 WS-13/JB01 biofilm was significantly greater
than that of a biofilm of JB01 alone (P<0?0001). In both
the WS-4/WS-5 and WS-13/JB01 tests, the increased MDM
of the mixed biofilms is not the result of differences in
growth rates, as in both cases, significant complementation was only seen with higher initial ratios of the strain
that could not produce a biofilm alone (i.e. the significant
comparisons are between 1 : 9 and 1 : 4 WS-4/WS-5, and
between 1 : 9 and 1 : 1 WS-13/JB01). These findings strongly
suggest that the strength of the WS biofilm is due to the
combination of cellulose, attachment factor and LPS,
and that partial complementation can be achieved by
expressing all three components by two strains in different
combinations.
3.9
0.015
0.010
0.005
WS-4 WS-5 1: 9
1: 4
9 : 1 WS-13 JB01 1: 99 1: 9
WS-4 : WS-5
1: 1
WS
WS-13 : JB01
Fig. 6. WS mutant A–L biofilms can be partially complemented
and show significant increases in strength. The WS-5 (WS
tol : : mini-Tn5) biofilm can be complemented by WS-4 (WS
wspR : : mini-Tn5), leading to a 1?76 increase in maximum
deformation mass (MDM) (white bars). WS-5 expresses attachment factor and cellulose, but not LPS; WS-4 expresses LPS,
but not cellulose or attachment factor. Similarly, the JB01 (SM
NPTII : : wss) biofilm is complemented by WS-13 (WS
wssB : : mini-Tn5), leading to a 3?96 increase in MDM (grey
bars). JB01 expresses cellulose and LPS, but not attachment
factor; WS-13 expresses LPS and attachment factor, but not
cellulose. Neither WS-4 nor WS-13 produces biofilms. The
MDM of WS in this experiment was 0?151±0?93 g (dark bar
off-scale). Microcosms were inoculated with single strains, or
with mixtures of strains, such that each test used the same
total number of cells. Microcosms were incubated at 20–22 6C
for 3 days before assay. Means±SE are shown.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
Microbiology 151
Involvement of LPS in the wrinkly spreader biofilm
DISCUSSION
In previous analyses of factors required for development
of the WS A–L biofilm, we demonstrated that partially
acetylated cellulose and a fimbrial-like attachment factor
were significant determinants of the strength and structural
integrity of the WS biofilm (Spiers et al., 2003). In this work,
we have extended our analysis to the wrinkly spreaderdefective mini-Tn5 mutant WS-5, which produces weak
biofilms despite the production of partially acetylated
cellulose and fimbrial attachment factor.
The site of the mini-Tn5 insertion in WS-5 was shown to be
immediately upstream of ybgC, the first gene in the highly
conserved tol-pal gene cluster, and it exerts polar effects on
downstream gene expression. The Tol-Pal system proteins
are involved in maintaining the correct functional relationship between the inner and outer membranes, and mutants
often show a variety of pleiotropic effects, including the loss
of LPS expression (Gaspar et al., 2000). We confirmed that
this was the case in WS-5 by DOC-PAGE and ELISA assays.
LPS plays a major role in determining the surface-charge or
relative hydrophobicity (Hr) of the cell (Rocchetta et al.,
1999), and is involved in bacterial attachment to surfaces
and biofilm formation; it differentiates between biofilm and
planktonic cells, as well as affecting colony morphology in a
number of bacteria (Giwercman et al., 1992; Genevaux et al.,
1999; Mireles et al., 2001; Nesper et al., 2001; Landini &
Zehnder, 2002; de Lima Pimenta et al., 2003; Rashid et al.,
2003). WS-5 cells showed a significantly different Hr from
WS cells, and maintenance of WS-5 cells in the biofilm was
less efficient than that of WS. In addition, the strength of the
WS biofilm was more sensitive to chemical interference than
that of the WS-5 biofilm. Each of these findings, along with
the demonstration that the WS-5 biofilm can be partially
complemented by an LPS-expressing strain, strongly suggests that LPS-dependent and charge-sensitive interactions
are important in WS biofilm development, and that complex interactions between cellulose fibres, attachment factor
and LPS determine the final strength of WS biofilms.
We have previously noted that the P. fluorescens WS A–L
biofilm and colony morphology are similar to those produced by Escherichia coli and Salmonella sp. (Spiers et al.,
2002), and in each case, cellulose fibres and curli/Tafi fimbriae are required for both biofilm strength and the rdar
colony morphology (Römling & Rohde, 1999; Zogaj et al.,
2001; Solano et al., 2002). In S. enterica biofilms, Tafi fibres
appear as a tangled amorphous matrix when cellulose is
present, but when it is not, the fibres adopt a more normal,
slightly curled linear structure (White et al., 2003). Further
analysis of E. coli and S. enterica biofilms has also revealed
the presence of a third matrix component, an anionic
extracellular polysaccharide, which requires cellulose in
order to maintain a close association with cells (White et al.,
2003). These findings, along with our observations regarding the WS biofilm, indicate that the structure and physical
properties of bacterial biofilms are the result of multiple
http://mic.sgmjournals.org
interactions between various matrix components – the main
EPS matrix fibres, proteinaceous attachment fibres (fimbriae and flagella), LPS and additional polysaccharides. In
parallel research, we are undertaking a comprehensive
screen for new WS mutants, in which ISphoA/hah disruptions of the cellulose acetylation genes, putative LPS
biosynthesis and membrane-associated genes have been
identified (S. Gehrig, A. Spiers & P. Rainey, work in progress), further underlining the importance of these interactions in the WS phenotype.
LPS is generally anchored to the bacterial outer membrane
via lipid A (Rocchetta et al., 1999). However, LPS is also
known to be released from cells during normal growth, and
cell-free LPS accumulates in cultures after cell lysis (Cadieux
et al., 1983; Ishiguro et al., 1986; Al-Tahhan et al., 2000).
This suggests that some of the cellulose fibre–attachment
factor–LPS interactions important to WS biofilm strength
may be cell-independent, insofar that once LPS has been
produced and released at one location, bacterial cells may
not be required to remain in place to maintain biofilm
strength. Indeed, the WS biofilm might result from the
aggregation of locally expressed and largely cell-free cellulose, attachment factor and released LPS, with the bacterial
cells free to move within the biofilm as it develops and as
environmental conditions change. This possibility is
supported by WS biofilm microscopy, in which few bacteria
were found to be closely associated with the cellulose matrix,
and most found to be mobile within the spaces of the biofilm
(Spiers et al., 2003).
Biofilm matrices are known to be chemically complex, with
85–98 % of the total organic carbon present as excreted
polymers and products from cell lysis, and the balance in
intact cells (Sutherland, 2001a). It is becoming increasingly
apparent that the physical structure of many biofilms is
not primarily the result of the expression of one matrix
component, but of several interacting elements. Furthermore, although the main component might be specifically
expressed during biofilm development, the expression of
other components may not be restricted to the biofilm. The
added complexity of matrix components and expression
patterns has an obvious impact on the study of biofilm
development. While the formation of biofilms may confer a
growth advantage at surfaces, biofilm matrix compounds
need not be biofilm-specific; for example, flagella and pili
play a role in motility as well as initial attachment, and EPS
glycocalyx may play a protective role as well as contribute
to formation of the biofilm matrix.
The finding that the structural aspect of biofilm matrices
is more complex than originally thought has implications for the ‘architectural’ nature of biofilm development
(Wimpenny et al., 2000; Ghigo, 2003). If biofilm development is a genetically programmed growth mode, then far
more biosynthetic pathways would need to be controlled
than currently recognized. On the other hand, if biofilm
growth is a consequence of bacterial attachment, the resulting community structure may be better defined by a small
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
2837
A. J. Spiers and P. B. Rainey
number of specific biosynthetic pathways, plus a number of
less-specific systems that contribute to the overall physicalchemical structure of the biofilm matrix. If this is so, then we
predict that biofilm growth will show a variable requirement for secondary matrix components and a degree of
redundancy, as these will not be uniquely required and may
be complemented by other components.
In the case of natural biofilms, ranging from those growing on dental surfaces or other human tissues to true water-,
soil- or plant-associated environmental biofilms (Davey &
O’Toole, 2000; Wilson, 2001; Morris & Monier, 2003),
different members of the biofilm may contribute to different
parts of the biofilm matrix. Given the continuously changing composition of biofilms during establishment, growth
and maturity, a dynamic and complex physical-chemical
matrix can only be expected. Many of the phenotypes associated with biofilms, such as increased resistance to stress
and antimicrobial agents, may be the consequence of the
microenvironment heterogeneity within biofilms (Ghigo,
2003). The complexity of both the biofilm matrix and
community underlies the ecological success of this type of
assemblage, and may also explain the value of such structures to both pathogenic and opportunistic bacteria.
Dunne, W. M. (2002). Bacterial adhesion: seen any good biofilms
lately? Clin Microbiol Rev 15, 155–166.
Gaspar, J. A., Thomas, J. A., Marolda, C. L. & Valvano, M. A. (2000).
Surface expression of O-specific lipopolysaccharide in Escherichia coli
requires the function of the TolA protein. Mol Microbiol 38, 262–275.
Genevaux, P., Bauda, P., DuBow, M. S. & Oudega, B. (1999).
Identification of Tn10 insertions in the rfaG, rfaP, and galU
genes involved in lipopolysaccharide core biosynthesis that affect
Escherichia coli adhesion. Arch Microbiol 172, 1–8.
Ghigo, J.-M. (2003). Are there biofilm-specific physiological path-
ways beyond a reasonable doubt? Res Microbiol 154, 1–8.
Giwercman, B., Fomsgaard, A., Mansa, B. & Hoiby, N. (1992).
Polyacrylamide gel electrophoresis analysis of lipopolysaccharide
from Pseudomonas aeruginosa growing planktonically and as biofilm.
FEMS Microbiol Immunol 4, 225–229.
Götz, F. (2002). Staphylococcus and biofilms. Mol Microbiol 43,
1367–1378.
Goymer, P. J. (2002). The role of the WspR response regulator in the
adaptive evolution of experimental populations of Pseudomonas
fluorescens SBW25. DPhil thesis, University of Oxford.
Groisman, E. A., Kayser, J. & Soncini, F. C. (1997). Regulation of
polymyxin resistance and adaptation to low-Mg2+ environments.
J Bacteriol 179, 7040–7045.
Hall-Stoodley, L. & Stoodley, P. (2002). Developmental regulation of
microbial biofilms. Curr Opin Biotechnol 13, 228–233.
Hall-Stoodley, L., Costerton, J. W. & Stoodley, P. (2004). Bacterial
biofilms: survival and propagation on surfaces from the environment
to infectious diseases. Nat Rev Microbiol 2, 95–108.
ACKNOWLEDGEMENTS
We thank J. Stansfield for her technical assistance and A. Whiteley for
his help in the FACS analyses. mAb BC12-CA4 was kindly made
available by M. Dewey. DNA sequences obtained in this work were used
to interrogate the unfinished P. fluorescens SBW25 genome sequence
(Wellcome Trust Sanger Institute, UK) to aid gene identification.
Funding for this work came in part from the BBSRC (UK).
REFERENCES
Ishiguro, E. E., Vanderwel, D. & Kusser, W. (1986). Control of
lipopolysaccharide biosynthesis and release by Escherichia coli and
Salmonella typhimurium. J Bacteriol 168, 328–333.
Kai, A. & Mondal, I. H. (1997). Influence of substituent of direct
dye having bisphenylenebis(azo) skeletal structure on structure of
nascent cellulose produced by Acetobacter xylinum [I]: different
influence of Direct Red 28, Blue 1 and 15 on nascent structure. Int
J Biol Macromol 20, 221–231.
King, E. O., Ward, M. K. & Raney, D. C. (1954). Two simple media for
Al-Tahhan, R. A., Sandrin, T. R., Bodour, A. A. & Maier, R. M.
(2000). Rhamnolipid-induced removal of lipopolysaccharide from
the demonstration of pyocyanin and fluorescin. J Lab Clin Med 44,
301–307.
Pseudomonas aeruginosa: effect on cell surface properties and
interaction with hydrophobic substrates. Appl Environ Microbiol
66, 3262–3268.
Landini, P. & Zehnder, A. J. (2002). The global regulatory hns gene
Cadieux, J. E., Kuzio, J., Milazzo, F. H. & Kropinski, A. M. (1983).
Spontaneous release of lipopolysaccharide by Pseudomonas aeruginosa. J Bacteriol 155, 817–825.
Costerton, J. W., Lewandowski, Z., Cladwell, D. E., Korber, D. R. &
Lappin-Scott, H. M. (1995). Microbial biofilms. Annu Rev Microbiol
49, 711–745.
Dalton, H. M. & March, P. E. (1998). Molecular genetics of bacterial
attachment and biofouling. Curr Opin Biotechnol 9, 252–255.
Davey, M. E. & O’Toole, G. A. (2000). Microbial biofilms: from
negatively affects adhesion to solid surfaces by anaerobically grown
Escherichia coli by modulating expression of flagella genes and
lipopolysaccharide production. J Bacteriol 184, 1522–1529.
Lappin-Scott, H. M. & Bass, C. (2001). Biofilm formation: attach-
ment, growth and detachment of microbes from surfaces. Am J Infect
Control 29, 250–261.
Lazzaroni, J. C., Germon, P., Ray, M. C. & Vianney, A. (1999). The
Tol proteins of Escherichia coli and their involvement in the uptake
of biomolecules and outer membrane stability. FEMS Microbiol Lett
177, 191–197.
ecology to molecular genetics. Microbiol Mol Biol Rev 64, 847–867.
Llamas, M. A., Rodrı́guez-Herva, J. J., Hancock, R. E. W., Bitter, W.,
Tommassen, J. & Ramos, J. L. (2003). Role of Pseudomonas putida
De Lima Pimenta, A., Di Martino, P., Le Bouder, E., Hulen, C. &
Blight, M. A. (2003). In vitro identification of two adherence factors
tol-oprL gene products in uptake of solutes through the cytoplasmic
membrane. J Bacteriol 185, 4707–4716.
required for in vivo virulence of Pseudomonas fluorescens. Microbes
Infect 13, 1177–1187.
Lloubes, R., Cascales, E., Walburger, A., Bouveret, E., Lazdunski, C.,
Bernadac, A. & Journet, L. (2001). The Tol-Pal proteins of the
Dennis, J. J. & Zylstra, G. J. (1998). Improved antibiotic-resistance
Escherichia coli cell envelope: an energized system required for outer
membrane integrity? Res Microbiol 152, 523–529.
cassettes through restriction site elimination using Pfu DNA polymerase PCR. Biotechniques 25, 772–776.
Donlan, R. M. (2002). Biofilms: microbial life on surfaces. Emerg
Infect Dis 8, 881–890.
2838
Meyer, U. & Dewey, F. M. (2000). Efficacy of different immunogens
for raising monoclonal antibodies to Botrytis cinerea. Mycol Res 104,
979–987.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
Microbiology 151
Involvement of LPS in the wrinkly spreader biofilm
Mireles, J. R., Toguchi, A. & Harshey, R. M. (2001). Salmonella
Simon, R., Priefer, U. & Puhler, A. (1983). A broad host range
enterica serovar typhimurium swarming mutants with altered
biofilm-forming abilities: surfactin inhibits biofilm formation.
J Bacteriol 183, 5848–5854.
mobilisation system for in vivo genetic engineering: random and
site-specific transposon mutagenesis in gram-negative bacteria.
Biotechnology 1, 784–791.
Morris, C. E. & Monier, J.-M. (2003). The ecological signifcance of
Solano, C., Garcia, B., Valle, J., Berasain, C., Ghigo, J.-M., Gamazo,
C. & Lasa, I. (2002). Genetic analysis of Salmonella enteritidis biofilm
biofilm formation by plant-associated bacteria. Annu Rev Phytopathol
41, 429–453.
formation: critical role of cellulose. Mol Microbiol 43, 793–808.
Nesper, J., Lauriano, C. M., Klose, K. E., Kapfhammer, D., Kraiss,
A. & Reidl, J. (2001). Characterization of Vibrio cholerae O1 El
Spiers, A. J., Kahn, S. G., Travisano, M., Bohannon, J. & Rainey,
P. B. (2002). Adaptive divergence in experimental populations of
tor galU and galE mutants: influence on lipopolysaccharide
structure, colonization, and biofilm formation. Infect Immun 69,
435–445.
Pseudomonas fluorescens. 1. Genetic and phenotypic bases of wrinkly
spreader fitness. Genetics 161, 33–46.
Okamura-Ikeda, K., Ohmura, Y., Fujiwara, K. & Motokawa, Y.
(1993). Cloning and nucleotide sequence of the gcv operon encoding
the Escherichia coli glycine-cleavage system. Eur J Biochem 216,
539–548.
Spiers, A. J., Bohannon, J., Gehrig, S. & Rainey, P. B. (2003).
Colonisation of the air-liquid interface by the Pseudomonas
fluorescens SBW25 wrinkly spreader requires an acetylated form of
cellulose. Mol Microbiol 50, 15–27.
Stoodley, P., Sauer, K., Davies, D. G. & Costerton, J. W. (2002).
plant rhizosphere. Environ Microbiol 1, 243–257.
Biofilms as complex differentiated communities. Annu Rev Microbiol
56, 187–209.
Rainey, P. B. & Bailey, M. J. (1996). Physical map of the
Sturgis, J. N. (2001). Organisation and evolution of the tol-pal gene
Pseudomonas fluorescens SBW25 chromosome. Mol Microbiol 19,
521–533.
Sutherland, I. W. (2001a). The biofilm matrix – an immobilized but
Rainey, P. B. & Travisano, M. (1998). Adaptive radiation in a
dynamic microbial environment. Trends Microbiol 9, 222–227.
heterogeneous environment. Nature 394, 69–72.
Rainey, P. B. & Rainey, K. (2003). Evolution of cooperation and
Sutherland, I. W. (2001b). Biofilm exopolysaccharides: a strong and
sticky framework. Microbiology 147, 3–9.
conflict in experimental bacterial populations. Nature 425, 72–74.
Van der Mei, H. C. & Busscher, H. J. (2001). Electrophoretic
Rashid, M. H., Rajanna, C., Ali, A. & Karaolis, D. K. (2003).
mobility distributions of single-strain microbial populations. Appl
Environ Microbiol 67, 491–494.
Rainey, P. B. (1999). Adaptation of Pseudomonas fluorescens to the
cluster. J Mol Microbiol Biotechnol 3, 113–122.
Identification of genes involved in the switch between the smooth
and rugose phenotypes of Vibrio cholerae. FEMS Microbiol Lett 227,
113–119.
White, A. P., Gibson, D. L., Collinson, S. K., Banser, P. A. & Kay,
W. W. (2003). Extracellular polysaccharides associated with thin
Reuhs, B. L., Geller, D. P., Kim, J. S., Fox, J. E., Kolli, V. S. K. &
Pueppke, S. G. (1998). Sinorhizobium fredii and Sinorhizobium
aggregative fimbriae of Salmonella enterica serovar enteritidis.
J Bacteriol 185, 5398–5407.
meliloti produce structurally conserved lipopolysaccharides and
strain-specific K antigens. Appl Environ Microbiol 64, 4930–4938.
235–254.
Rocchetta, H. L., Burrows, L. L. & Lam, J. S. (1999). Genetics of O-
Wimpenny, J., Manz, W. & Szewzyk, U. (2000). Heterogeneity in
antigen biosynthesis in Pseudomonas aeruginosa. Microbiol Mol Biol
Rev 63, 523–553.
biofilms. FEMS Microbiol Rev 24, 661–671.
Römling, U. & Rohde, M. (1999). Flagella modulate the multicellular
behaviour of Salmonella typhimurium on the community level. FEMS
Microbiol Lett 180, 91–102.
http://mic.sgmjournals.org
Wilson, M. (2001). Bacterial biofilms and human disease. Sci Prog 84,
Zogaj, X., Nimtz, M., Rohde, M., Bokranz, W. & Römling, U. (2001).
The multicellular morphotypes of Salmonella typhimurium and
Escherichia coli produce cellulose as the second component of the
extracellular matrix. Mol Microbiol 39, 1452–1463.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:18:29
2839