REVIEWS

UNDERSTANDING BIOFILM
RESISTANCE TO ANTIBACTERIAL
AGENTS
David Davies
According to a public announcement by the US National Institutes of Health, ‘‘Biofilms are
medically important, accounting for over 80% of microbial infections in the body’’. Yet bacterial
biofilms remain poorly understood and strategies for their control remain underdeveloped.
Standard antimicrobial treatments typically fail to eradicate biofilms, which can result in chronic
infection and the need for surgical removal of afflicted areas. The need to create effective
therapies to counter biofilm infections presents one of the most pressing challenges in anti-
bacterial drug development. In this article, the mechanisms that underlie biofilm resistance to
antimicrobial chemotherapy will be examined, with particular attention being given to potential
avenues for the effective treatment of biofilms.

CONFOCAL LASER SCANNING A biofilm is a population or community of bacteria Knockout experiments have shown that the genes
MICROSCOPY living in organized structures at a liquid interface. that control the biosynthesis of these messengers are
A microscopy technique which Early CONFOCAL LASER SCANNING MICROSCOPY (CLSM) of involved in the the formation of a differentiated
uses scanning laser light to
single-species biofilms1,2 revealed that biofilm bacteria Pseudomonas aeruginosa biofilm, and the knockout
excite fluorescent dyes within a
thick sample, such as a biofilm. live in cellular clusters or MICROCOLONIES that are encapsu- phenotype develops a biofilm of densely packed cells
The image is collected in two lated in a matrix composed of an extracellular polymeric a few layers thick5. A representative diagram of a
dimensions and several images substance (EPS), separated by open water channels that biofilm formed by the pathogen P. aeruginosa is
can be combined in an image act as a primitive circulatory system for the delivery of depicted in FIG. 1.
stack to produce a cross
sectional image through a
nutrients and the removal of metabolic waste products.
sample or to create a three- Fluid flow within the water channels was discovered Biofilms and human disease
dimensional rendering of the using nuclear magnetic resonance (NMR) imaging Biofilms have been increasingly recognized as being
sample. CLSM is particularly and serial CLSM observations of inert particles; the important in human disease. The number of diseases
useful for imaging the
rate and direction of this flow having been determined associated with bacterial biofilms is considered to be
positioning of biological
structures within a three in several systems3,4. quite large, with colitis, vaginitis, urethritis, conjunctivitis
dimensional space. Within a biofilm, each bacterium occupies a specific and otitis being just a short list of common examples.
microenvironment, which is determined by surround- Biofilm infections have been known to be problematic
ing cells, proximity to a channel (both of which deter- in the oral cavity, and GINGIVITIS serves as an example of
mine the pH and availability of nutrients and oxygen) the prevalence of such infections. It has been reported
Department of Biological and the EPS matrix. The structuring of biofilms in that 24% of adults have lost at least 4 mm of periodontal
Sciences, State University microcolonies and water channels has been shown to be attachment, and 60% of 15-year-olds and 40–50% of
of New York, Binghamton, influenced by fluid flow, nutrient composition and adults have some form of gingival (biofilm) infection8,9.
New York 13902, USA.
e-mail:
intercellular small messenger molecules, or quoro- Biofilms are also important as colonizers of medical
[email protected] mones (acylated homoserine lactones, AHLs), that are devices, including urinary, venous and arterial
doi:10.1038/nrd1008 used for bacterial communication5-7 (see BOX 1). catheters10 and shunts. In a study of 4,000 infants given

114 | FEBRUARY 2003 | VOLUME 2 www.nature.com/reviews/drugdisc

 REVIEWS

Box 1 | Quorum sensing Although numbers are difficult to evaluate critically, it is

taken for granted within the biofilm research commu-
Recent advances in research on cell–cell communication in bacteria have demonstrated nity that many undiagnosed chronic diseases in humans
a roll for chemical signalling in bacterial biofilms. This research has shown that small, might be of biofilm origin.
diffusible molecules — members of the class of N-acylated homoserine lactones (AHLs) Problems associated with human biofilm infections
— are released by biofilm bacteria into their local environment, where they can interact result from two distinct characteristics of all biofilms.
with neighbouring cells. In all cases, AHLs are known to associate with a cognate First, biofilms are highly resistant to immune killing and
DNA-binding protein that is homologous to LuxR in Photobacterium fischeri, which
clearance, and to treatment with antimicrobial agents12,13.
causes a conformational change in the protein that facilitates DNA-polymerase binding
Second, protected biofilms might be capable of shedding
and initiates transcription of target genes. As bacterial densities increase with growth,
individual bacteria and sloughed pieces of biofilm into
these AHLs can accumulate to a threshold concentration and induce the transcription
surrounding tissues and the circulatory system. Such
of specific genes throughout the population. This process couples the transcription of
specific genes to bacterial-cell density89. Regulation of this type has been referred to as shed cells might be responsible for acute illness, which
‘quorum sensing’, because it suggests the requirement for a ‘quorate’ population of might recur despite vigorous antimicrobial treatments.
bacterial cells that is necessary for the activation of AHL-responsive genes90. Regulation
of this type enables the coordination of bacterial behaviour at the population level, and Biofilm resistance to antimicrobial agents
ensures that the bacteria respond as a group to carry out special functions. For instance, Owing to the compact nature of biofilm structures, the
quorum sensing has been shown to be responsible for the release of degradative presumed reduced rates of cellular growth and respira-
extracellular enzymes and cytotoxins in a number of bacterial species. It is advantageous tion of biofilm bacteria and the protection conferred by
for bacteria to act as a group, rather than as autonomous individuals. The role of quorum biofilm matrix polymers, natural and artificial chemical
sensing in bacterial infections is still incompletely characterized; however, great effort is agents are unable to adequately attack and destroy
currently being put into the investigation of this phenomenon. These studies are expected infectious biofilm populations10,11. Increased antibiotic
to yield important dividends in the development of new anti-infective chemotherapies. resistance is a general trait associated with biofilm
bacteria. When attached, bacteria show a profound
resistance, rendering biofilm cells 10–1,000-fold less sus-
cerebrospinal-fluid shunts, 15–20% were found to be ceptible to various antimicrobial agents than the same
infected by a biofilm11. In addition, respirators, sigmoi- bacterium grown in PLANKTONIC (free-floating) culture. For
doscopes, contact lenses, artificial implants (for example, instance, chlorine (as sodium hypochlorite) — an oxidiz-
heart valves, pacemakers, ventricular assist devices, syn- ing biocide that is considered to be one of the most
thetic vascular grafts and stents), urinary prostheses and effective antibacterial agents — requires a 600-fold
orthopaedic prostheses (such as artificial joints and pins), increase in concentration to kill biofilm cells of
have all been shown to be infected with biofilms. It has Staphylococcus aureus compared with planktonic cells
even been speculated that breast implantation-associated of the same species14. Several factors have been sug-
medical problems might arise primarily from biofilm gested to account for the extraordinary resistance of
infections on the implant material rather than from the biofilm bacteria to antibiotics: the reduced metabolic
implant itself (G. Ehrlich, personal communication). and growth rates shown by biofilm bacteria, particularly

MICROCOLONY
A microscopic aggregation of
cells in a biofilm.

GINGIVITIS
Infection of the gingival crevice
(periodontal pocket) of the oral
cavity with a variety of
microorganisms, causing
inflammation of the periodontal
tissue and bone loss. Caused by
members of the genus
Capnocytophaga, Porphyromonas,
Rothia and others.
Figure 1 | Five stages of biofilm development. Biofilm maturation is a complex developmental process that involves several
PLANKTONIC stages, each with unique characteristics that should be considered when designing strategies for biofilm treatment with antibiotics.
Organisms that are free-floating Each stage of development in the diagram is paired with a photomicrograph of a developing Pseudomonas aeruginosa biofilm.
in a fluid environment. All photomicrographs are shown at the same scale. Modified with permission from REF. 58 © (2002) American Society for Microbiology.

NATURE REVIEWS | DRUG DISCOVERY VOLUME 2 | FEBRUARY 2003 | 1 1 5

REVIEWS

Untreated 20 minutes 24 hours

Live Killed

Figure 2 | Biofilm resistance to anibiotic addition. Treatment of biofilms with antibiotics often results in incomplete killing,
allowing unaffected bacteria to act as a nucleus for the spread of infection following the withdrawal of antibiotic therapy.

those deep within the biofilm, might make them inher- gentamicin, tilmicosin and trimethoprim-sulphadoxine.
ently less susceptible to antibiotics; the biofilm EPS All of these antibiotics were effective in the treatment of
matrix might act as an adsorbent or reactant, thereby planktonic cultures of Actinomyces pyogenes, Coryne-
reducing the amount of agent available to interact with bacterium renale, C. pseudotuberculosis, Staphylococcus
biofilm cells (additionally, the biofilm structure might aureus, S. hyicus and Streptococcus agalactiae ; however,
physically reduce the penetration of antimicrobial biofilms formed by all of these organisms were resistant
agents by walling off access to regions of the biofilm); to all of the antibiotics tested. Some bacteria were
and biofilm cells are physiologically distinct from shown to have biofilm susceptibilities that were equal
planktonic bacteria, and express specific protective to planktonic cultures, including Pasteurella multocida,
factors, such as MULTIDRUG EFFLUX PUMPS and stress- Mannheimia haemolytica, Streptococcus suis and
response REGULONS15–22. As detailed molecular studies S. dysgalactiae. Salmonella sp. and Pseudomonas
emerge, it is becoming apparent that each of these factors aeruginosa were susceptible to enrofloxacin, gentam-
is important in the unusual resistance of biofilms to icin, ampicillin, oxytetracycline and trimethoprim-
antimicrobials. FIGURE 2 represents the activity of sulphadoxine as planktonic cultures, but were sensitive
antibiotics against a typical biofilm population. Initial as biofilms only to enrofloxacin.
treatment is usually effective in killing bacteria only at Biofilms also show enhanced resistance to host-
the margins of biofilm microcolonies. Bacteria deep defence mechanisms. In chronic infections, such as the
within these microcolonies are not always killed by the bronchopulmonary P. aeruginosa infection in cystic
antibacterial agents23,24, and can potentially form a NIDUS fibrosis (CF) patients (see BOX 2), bacteria persist despite
MULTIDRUG EFFLUX PUMP
for continued dissemination of the infection. an intact host immune defence and frequent antibiotic
A molecular pump integrated
into the cell envelop of certain There are growing concerns that the antibiotic treat- treatment. An important reason for the persistence of
bacteria which is able to ment of NOSOCOMIAL infections and wound infections is the bacteria is their capacity for the biofilm mode of
transport antibiotics into and driving the evolution of antibiotic-resistant micro- growth. Biofilm-grown P. aeruginosa showed reduced
out of the cell. organisms, and concerns regarding the over-prescription activation of COMPLEMENT compared with planktonic
REGULON
of antibiotics are increasingly being raised25,26. These bacteria29. Furthermore, the aggregation of bacteria into
A set of operons that are concerns are reflected in a publication by Rastegar et al.27, EPS-coated biofilms might make them less susceptible
controlled by a single which reported the identification of P. aeruginosa as the to phagocytosis30–33. Biofilm bacteria have also been
regulatory protein. most common causes of wound infection in burn reported to be resistant to certain aspects of the HUMORAL
29,34,35
patients, with a frequency of 73.9%. The frequency of P. IMMUNE SYSTEM . The accessibility is presumed to be
NIDUS
Latin for nest, but in this context aeruginosa that is resistant to gentamicin, carbenicillin, reduced due to the protective EPS. Persistence of
a place or point in a host where a co-trimoxazole, ceftizoxime and tetracycline was biofilms results in damage to the host as phagocytic cells
pathogen can develop and breed. more than 95%. These results exemplify the need to will release indiscriminately their oxidative burst, result-
develop alternate anti-infective strategies, including ing in collateral tissue damage.
NOSOCOMIAL
Something acquired or
new antibiotics and vaccines, which can be used
originating in a hospital, such as against P. aeruginosa and other pathogens. Antibiotic penetration
a nosocomial infection. The resistance of biofilms to antibiotic treatments One of the factors that is generally conceded to have a
depends on the microorganism under consideration and role in antibiotic resistance by biofilms is the inability of
COMPELEMENT
the antibiotic used. In a study by Olson and colleagues28, the antibiotic to penetrate to all areas of the biofilm.
A complex of blood serum
proteins of the immune system the Calgary Biofilm Device (a continuous-flow device Several studies have been carried out in which antibiotic
that interact sequentially with that is used for the culture of biofilms) was used to assess penetration has been assessed by detecting the concen-
antibody–antigen complexes. the susceptibility of planktonic versus biofilm cultures of tration of the antibiotic at the base of the biofilm. In one
a number of different bacterial species and several differ- such series of experiments, the penetration of the
HUMORAL IMMUNE SYSTEM
Extracellular branch of the
ent antibiotics. The antibiotics tested included ampi- antibiotic ciprofloxicin was investigated for its ability to
immune system mediated by cillin, ceftiofur, cloxacillin, oxytetracycline, penicillin G, pass through biofilms of P. aeruginosa to a germanium
antibodies. streptomycin, tetracycline, enrofloxacin, erythromycin, crystal substratum in an infrared (IR) field. Germanium

116 | FEBRUARY 2003 | VOLUME 2 www.nature.com/reviews/drugdisc

 REVIEWS

MINIMUM INHIBITORY
crystal is transparent to IR radiation which passes, tetracycline penetration into biofilms formed by
CONCENTRATION unimpeded, through the crystal to create an evanescent Escherichia coli. In this study, it was shown that biofilms
The minimum concentration of field extending 0.2 µm above the surface. The IR signa- formed over two days on a polystyrene surface were less
a substance required to prevent ture of a material (such as an antibiotic) that is located susceptible to the antibiotic than were planktonic cells;
growth of a microoganism.
within the evanescent field is, therefore, detectable and however, the biofilms showed tetracycline-mediated
EXOPOLYMERIC MATRIX can be monitored in situ and in real time. Results from fluorescence distributed throughout the entire biofilm
A network of long-chain these experiments established that the biofilm was able following exposure to the antibiotic for 7.5–10 minutes39.
polymers produced by to significantly reduce, but not entirely block, antibiotic Although this study did not provide quantitative data on
microorganisms of a biofilm
penetration36. These and subsequent results also showed the concentration of tetracycline within the biofilm, it
which supports the structure
of the biofilm. that penetration rates through biofilms depended on nonetheless demonstrated that the antibiotic was able to
the antibiotic used and are not directly correlated with penetrate to all observable areas of the biofilm.
antibiotic susceptibility 37. Biofilm populations are typically found at markedly
Wild-type Klebsiella pneumoniae grown on filter higher cell densities than planktonic bacteria. It has been
discs were shown to have reduced antibiotic penetration noted that the higher cell density of a biofilm popula-
for ampicillin compared with ciprofloxicin. However, tion might account, in part, for their enhanced resis-
β-lactamase-deficient K. pneumoniae biofilms — in tance to antimicrobial treatment. Larsen40 has reported
which ampicillin was shown to penetrate completely — that the MINIMUM INHIBITORY CONCENTRATIONS (MICs) of the
were still resistant to treatment, with a log reduction of antibiotics amoxicillin, doxycycline and metronidazole
0.18 for the mutant strain, compared with 0.06 for the were all shown to increase markedly when planktonic
wild type and > 4 for the wild type in planktonic populations of Porphyromonas gingivalis were tested at
culture38. These results indicated that reduced antibiotic cell densities equal to those found in biofilm popula-
penetration might be important in protection for certain tions (107–108 cells ml–1). Although this result indicated
antibiotics, but that this reduction could not account for that an inoculum effect is part of the explanation for the
the overall resistance of biofilms to antibiotic treatment. increased resistance of biofilm bacteria, it does not com-
Although numerous studies have described the pene- pletely explain differences between biofilm and plank-
tration of antibiotics into biofilms, the manner in which tonic susceptibilities. Larsen also noted that biofilm
these studies were performed should be a consideration populations were still at least two to eight times more
when interpreting these results and in assessing their sig- resistant to amoxicillin and doxycycline compared to
nificance. It is typical to grow a biofilm on a permeable equivalent numbers of planktonic bacteria.
support or a substratum through which the antibiotic The presence of the EXOPOLYMERIC MATRIX of biofilms
can be detected directly. In such studies, the ability of the has long been held to have a role in limiting the penetra-
antibiotic to reach the substratum might be a function of tion of antimicrobials to cells deep within biofilms.
biofilm surface coverage in addition to biofilm penetra- Roques and colleagues41 have provided support for this
tion. FIGURE 3 illustrates this problem, and shows that asumption by showing that increasing the size of the
when antibiotic is added to a biofilm culture, it can pass hydrophobic side chains of selected quaternary ammo-
through gaps in the interstices between microcolonies nium compounds reduces the susceptibility of S. aureus
and lead investigators to conclude that the antibiotic to treatment with antibiotics when these bacteria are
concentration beneath the biofilm is equivalent to the embedded in a hydrophobic matrix of EPS. When the
concentration within the biofilm. Ideally, penetration matrix was washed from the cells, the susceptibility rose
studies should be performed by taking measurements at to 90% of the susceptibility of planktonic bacteria41.
the centre of microcolonies. Resistant biofilm bacteria become susceptible to anti-
In an effort to address the question of antibiotic pene- microbial treatments following dispersion or disaggre-
tration into dense cell aggregates of biofilms, Matin gation of the biofilm — an observation that further
and colleagues used direct microscopic observation of supports the idea that the EPS matrix might impart
protection to the biofilm by limiting transport (D. G.
Davies, unpublished observations).
One promising solution to the problem of antibiotic
Antibiotic penetration Antibiotic penetration
penetration has emerged from the manipulation of elec-
trical fields that surround bacteria in a biofilm. This
‘Bioelectric Effect’ — a term coined by J. W. Costerton
and colleagues — has been postulated to electrically
alter the configuration of the EPS matrix surrounding
biofilm bacteria, and perhaps also to enhance the pene-
tration of antimicrobial agents across the bacterial-cell
envelope42. Using alternating-current densities of less
Antibiotic penetrates Antibiotic penetrates
to substratum to substratum through than 100 micro-Amperes per cm2, it was found that the
through biofilm gaps within biofilm antibiotic concentrations required to kill biofilm cells
Figure 3 | Antibiotic penetration. Antibiotic penetration through biofilms is commonly
were greatly reduced compared with untreated biofilm
determined by how much antibiotic is detectable beneath the biofilm following addition of bacteria. These concentrations, however, were still
antibiotic. These measurements are influenced by the amount of biomass and percent higher than those needed to kill planktonic bacteria of
surface coverage by the biofilm. the same species43,44.

NATURE REVIEWS | DRUG DISCOVERY VOLUME 2 | FEBRUARY 2003 | 1 1 7

REVIEWS

AUXOTROPH
Reduced growth in biofilms The lowered metabolic activity of at least some
An organism that has acquired a Microbiologists have long known that non-dividing biofilms might therefore account for their enhanced
nutritional requirement through bacteria escape the killing effects of antibiotics targeted resistance to treatment with antibiotics that are active
the process of mutation. against growth-specific factors. For example, early peni- against growth factors in planktonic bacteria. Obser-
SPORULATION
cillin selection experiments to obtain E. coli AUXOTROPHS vations of the partial killing of bacteria within biofilms
The production an endospore by relied on the differential replication rates of wild-type and by such antibiotics can be explained by this type of
bacteria of the genera Clostridia nutrient-dependent organisms. Many investigators resistance. Biofilms that are treated with antibiotics can
and Bacillus. accept that biofilm bacteria have reduced growth rates, potentially sustain mortality on the periphery while the
and that this is a contributing factor to the unusual resis- deeper organisms persist and form a nucleus for re-
SIGMA FACTOR
Any of several bacterial DNA- tance of biofilms to effective antibiotic treatment.Yet evi- growth, and, through the use of metabolic stains for the
binding proteins that direct the dence for reduced growth in biofilms is limited and many detection of respiratory activity, this has been shown to
binding of DNA-directed RNA- reports of reduced activity arise primarily from unpub- be the case23,24. Cells at the bulk water interface of
polymerase to the promoter of lished observations. One reason for the dearth of direct biofilms are rapidly killed, whereas deep within cell
an operon.
information relating to divisional activity within biofilms clusters near the substratum bacteria remain active and
is the difficulty of making such measurements. Soren apparently unaffected by the treatment.
Molin’s research group at the Danish Technical University Antibiotics with activity against non-growing cells
has been one of the leaders in the study of metabolic have been shown to have enhanced activity against
activity within biofilms, and work in this lab using fluo- biofilm bacteria compared with antibiotics that do not
rescent tags for specific metabolic markers has shown that depend on rapid growth for activity. When imipenem
cells in the centres of the largest microcolonies in biofilms and ciprofloxicin were used on E. coli grown as a
do, in fact, have reduced metabolic rates compared with biofilm, their effectiveness was greater when compared
cells at or near the surface (FIG. 4). In some instances, cells with antibiotics (such as β-lactams) that were only
in small microcolonies also show reduced metabolic effective against growing bacteria, but less effective
activity. These observations indicate that nutrient avail- than the same antibiotics used against planktonic
ability is a crucial factor influencing metabolic activity E. coli 46. Similar results have been reported for
and, if properly supplied, cells within even large micro- P. aeruginosa, which, as a biofilm, shows greater sus-
colonies continue to be active at rates equivalent to those ceptibility to fluoroquinolone antibiotics compared
seen for planktonic cultures45. These researchers pointed with β-lactam antibiotics47–49.
out that during the initial phases of biofilm development, In another study, Spoering and Lewis50 compared
all cells at the colonization surface are highlyactive, with the slow growth of sub-populations of P. aeruginosa
ribosomal promoter activity corresponding to that of biofilm cells with stationary phase cells. The authors
rapidly growing cells. Following primary colonization reported that the antibiotic resistance of these two
and the formationof small microcolonies, activity gradu- populations was similar, and concluded that slow
ally decreases, initially inthe central parts of the micro- growth is potentially a factor in resistance. However,
colony and eventually also at the surface45. In related caution should be exercised when interpreting these
work, DeBeer et al.4 have directly measured oxygen con- results, as these data were not normalized and biofilm
centrations at various depths of biofilms using micro- cells and stationary-phase cells might have other attrib-
electrodes. This work has shown that the oxygen levelsare utes in common in addition to lowered metabolic rates,
depleted by as much as 30-fold near the centre of larger such as the activation of stress-response regulons and
microcolonies. This work indicates that other nutrients, the activation of efflux mechanisms.
such as organic carbon, will likewise be depleted towards
the centre of microcolonies. These studies have been per- Unique biofilm physiology
formed in continuous culture, but it is expected that the Research on biofilms has demonstrated differences
nutritional status of biofilm bacteria in an infection will between planktonic and attached bacteria that imply
have considerably less nutrient available and, as such, physiological alterations following attachment to
growth rates are expected to be correspondingly lower a surface. Early observations from a wide variety of
under these conditions. laboratories51–54 stimulated the development of the
hypothesis that biofilm bacteria were potentially physio-
logically distinct from planktonic bacteria.
Building on this earlier work, it has been established
High activity that biofilm bacteria display unique gene-expression
patterns, and, furthermore, that these patterns are not
Intermediate activity
observed in free-living bacterial cells18,55–59. The attach-
Low activity ment of bacteria to a surface initiates the expression of
Dormant biofilm-specific genes, culminating in what has been
described by Costerton33 as a ‘biofilm phenotype’. In
terms of bacterial gene-expression mechanisms, this
Figure 4 | Metabolic activity in a biofilm mirocolony. Metabolic activity in a biofilm cell cluster
major phenotypic change is analogous to SPORULATION or
is a function of depth within the biofilm and is influenced by nutrient transport. Cells at the edges starvation/survival. These phenotypic changes can
of a microcolony at the bulk liquid interface are the most active. Cells deep within the microcolony occur as a result of the induction of a series of RNA-
can potentially remain dormant and provide a nucleus for regrowth following antibiotic treatment. polymerase-associated SIGMA FACTORS60, which results in

118 | FEBRUARY 2003 | VOLUME 2 www.nature.com/reviews/drugdisc

 REVIEWS

attachment18. In work on Streptococcus mutans, Svensater

Box 2 | Biofilms and the cystic fibrosis lung
and colleagues76 detected increased concentrations for 57
Cystic fibrosis (CF) is the most prevalent lethal genetic disease among people of out of 694 proteins analysed in biofilm populations,
European descent. In the United States, approximately 30,000 children and adults are compared with planktonic populations. Thirteen pro-
afflicted. CF is inherited as an autosomal recessive trait at a rate of 1 in 2,000 live births teins in biofilm cells were not detected in planktonic
among Caucasians91. The hallmark of CF is the progressive loss of pulmonary function cultures and nine proteins were found only in planktonic
caused by chronic bacterial infection, typically with mucoid Pseudomonas aeruginosa. cultures. Chemostat-grown P. aeruginosa planktonic
In such infections, the bacteria persist despite an intact host immune defence and cells, compared with mature biofilm cells grown in sili-
frequent antibiotic treatment. An important reason for the persistence of the bacteria is con tubing in identical media, reveal more than 800 pro-
their capacity for the biofilm mode of growth32,33. Direct examination by confocal teins that have a sixfold or greater change in expression
scanning light microscopy (CSLM) of cystic fibrosis sputum has confirmed the presence level (more than 50% of the proteome). This difference
of biofilm-like structures encased in a polymeric matrix92. CSLM can be used on live
was higher than when planktonic P. aeruginosa were
specimens without altering the natural structure of the biofilm. Furthermore,
compared with planktonic cultures of P. putida58. These
extracellular quorum-sensing signalling molecules (extracellular chemical signals that
results indicate that physiological changes in the transi-
cue cell-density-dependent gene expression) were detected in sputum90 that were shown
tion from planktonic to attached cells are profound and
by Davies et al.5 to coordinate biofilm formation.
undoubtedly complex. An example of comparative two-
dimensional polyacrylamide gels (2D PAGE) for biofilm
and planktonic populations of P. aeruginosa is shown in
biofilm bacteria that are morphologically and biochemi- FIG. 5. These gels highlight the numerous differences in
cally distinct from their free-floating counterparts32. In a cellular protein profiles that give rise to the differences in
recent review, O’Toole et al.61 described biofilm forma- physiology between biofilm and planktonic forms of the
tion as a process of microbial development similar to same organism.
that seen in cell-cycle-controlled SWARMER-TO-STALK cell The differences in gene expression and protein pro-
transition in Caulobacter crescentus, sporulation in files that are seen between planktonic and biofilm popu-
Bacillus subtilis and FRUITING-BODY formation by lations indicate that several factors are likely to be
Myxococcus xanthus. From a structural point of view, it responsible for the differences observed in the resistance
is similar to the formation of tissue from a variety of to antibiotics by these two modes of bacterial growth.
individual cells. This view is gaining increasing accep- Recent studies have shown that the disruption of the
tance as studies on initial events in biofilm development expression of potential multidrug efflux pumps in
reveal alterations in bacterial-cell physiology that hint at biofilms of P. putida using KT 2410 TRANSPOSON insertion
changes that can occur throughout the developmental results in attachment-defective mutants77. This indicates
cycle55,58,62–74. The concept of a unique biofilm pheno- that several biofilm-associated traits are under unified
type is crucial to identifying new targets for controlling regulatory control and that numerous mechanisms,
bacterial infections. which are associated with antibiotic resistance, might be
Recent investigations have been directed at deter- operative at the same time within a biofilm.
mining the degree to which gene regulation during It is interesting to note that the bacteriacidal activity
SWARMER-TO-STALK CELL
TRANSITION
biofilm development controls the switch from plank- of biocides, such as chlorine and glutaraldehyde, cannot
Upon exhaustion of nutrients, tonic to biofilm growth. By monitoring changes in be adequately neutralized by specific resistance mecha-
members of the group of global gene-expression patterns in attached Pseudomonas nisms, such as efflux pumps, which are effective against
fruiting myxobacteria swarmer aeruginosa cells, Brözel and colleagues75 found that the antibiotics. Chlorine and glutaraldehyde are, therefore,
cells migrate together and
undergo differentiation into
expression levels of at least 11 proteins were altered able to kill biofilm bacteria if their concentration is suffi-
stalk cells, forming a vertical during various stages of attachment. Genevaux and ciently high. These biocides were tested against P. aerugi-
structure rising above a surface. colleagues71 screened a library of E. coli Tn10-insertion nosa embedded in calcium alginate beads, and it was
mutants with altered adhesion abilities: fifty adhesion- found that these ‘biofilm’ organisms were less suscepti-
FRUITING-BODY
deficient mutants were isolated that showed less than ble to treatment with chlorine and glutaraldehyde com-
A structure of the fruiting
myxobacteria at the end of a 40% attachment compared with the wild type, and 22 pared with planktonic bacteria, but complete killing was
stalk composed of differentiated mutants were found with an attachment of 40–75% achievable if the concentration of these biocides was suf-
cells which are converted to compared with the wild type. The majority of these ficiently high78. This indicates that the overriding factor
myxospores (resting bodies). mutants showed defects in motility. Using a screen in contributing to biofilm resistance to specific antibiotics
TRANSPOSON
E. coli K-12, similar to the approach of Genevaux et al.71, might be the specific genes that are activated (and deac-
A mobile segment of DNA that Prigent-Combaret and colleagues56 revealed major tivated) during biofilm growth, rather than nonspecific
has the ability to integrate into a changes in the patterns of gene expression during the mechanisms for protection, which should also provide
chromosome. Transposons switch from planktonic to attached growth. Attachment- protection against biocides such as chlorine.
usually carry genes that are used
dependent regulation of gene expression was seen in
in transposition as well as other
genes, often selectable markers, 38% of the generated lacZ gene fusions (out of 446 Biofilm detachment
such as for antibiotic resistance. clones). Using gene-chip technology, Whitely et al. The process of AUTODISPERSION or disaggregation of
detected 72 genes that showed differential expression in biofilm cells is of considerable interest as a means of
AUTODISPERSION biofilm populations of P. aeruginosa compared with increasing the antibiotic sensitivity of biofilm bacteria.
The disaggregation of a biofilm
or biofilm microcolony as a
planktonic cultures57. More recently, it has been shown The ability to induce dispersion has potential in control-
result of physiological activity that in Pseudomonas putida more than 30 genes and 40 ling biofilms directly; alternatively, it might be used as an
of the resident microorganisms. gene products were altered within 6 hours following adjuvant to existing antimicrobial therapies to enhance

NATURE REVIEWS | DRUG DISCOVERY VOLUME 2 | FEBRUARY 2003 | 1 1 9

REVIEWS

a b

Figure 5 | 2D PAGE gels of Pseudomonas aeruginosa. a | Protein profile from plantonic bacteria. b | Climax stage, six-day-old
biofilm. Modified with permission from REF. 58 © (2002) American Society for Microbiology.

their cidal activity. In 1998, Allison and co-workers one species but not in another 84–86. These results indicate
observed that spent medium from P. fluorescens cultures that hosts differ in their response to different virulence
was able to induce dispersion of biofilms formed by this factors and that positive results for mutants of bacteria
bacterium79; the same phenomenon has been observed attenuated in a particular virulence gene should not be
with biofilms formed by P. aeruginosa 5,58. Vats and Lee 80 automatically extrapolated to alternate hosts.
described the discovery that surface-protein-releasing Another area of pathogenesis research that requires
enzyme (SPRE) produced by Streptococcus mutans is additional attention considers which specific proteins
actively involved in the degradation of attachment poly- are produced during a biofilm infection. In a recent
mers on tooth surfaces, which causes the release of publication, Drenkard and Ausubel59 identified a regu-
bacteria from the tooth surface. In this work, SPRE was latory protein, PvrP, in P. aeruginosa that controls the
shown to result in a 20% increase in detachment com- conversion between antibiotic-resistant and susceptible
pared with control samples. In a more recent paper, forms. This gene has been shown to be actively tran-
Jackson et al.81 describe the discovery that the RNA- scribed by biofilm bacteria; compounds that inhibit the
binding protein CsrA (carbon storage regulator) acts as expression or activity of PvrP might prove to be useful
an activator of biofilm dispersal in E. coli. The effects of in treatment of P. aeruginosa biofilm infections. It is
CsrA are proposed to be mediated by regulation of expected that many bacteria will be found to produce
intracellular glycogen biosynthesis and catabolism. specific resistance or virulence gene products, which can
Biofilm dispersion is an almost untouched area of act as targets for anti-infective treatments.
research, but one which has the promise of providing
significant opportunities for alternative approaches to Future directions of biofilm research
biofilm control in the coming years. One of the main difficulties in treating biofilm infec-
tions arises from a lack of understanding of the char-
Targets for novel biofilm antibacterial agents acteristics of the biofilm mode of growth. So far, no
Recent studies on the pathogenesis of Pseudomonas infec- comprehensive investigation has taken place to examine
tions have begun to take advantage of presently available the phenotypic characteristics of the entire growth
technology to identify virulence genes and mRNA trans- cycle of a biofilm-forming bacterial pathogen.
cripts from infections in humans, mice and a number of Understanding when and how to treat biofilm infec-
animal models, including fruit flies, Caenorhabditis tions requires knowledge of whether the bacterial popu-
elegans and the greater wax moth Galleria mellonella82,83. lation is of one phenotype or of many. If separate
For instance, insect models of acute-stage infections of phenotypes are shown to exist, then antibiotic treat-
the HAEMOLYMPH by P. aeruginosa pathogenesis have been ments might not be effective against the entire infec-
generated, allowing the examinination of free-floating tion, leaving cells behind to recolonize and debilitate
bacteria (F. M. Ausubel, personal communication). the host. The presence of such sub-populations of cells
A P. aeruginosa mutant, defective in the gene ybtQ, was has been noted by Lewis and colleagues who refer to
shown to have attenuated virulence in both the wax moth these resistant bacteria as ‘persister’ cells50,87. A number
and the burned-mouse model, but not in C. elegans. of suggestions have been put forward to explain the
HAEMOLYMPH
The body fluid that bathes
Previous studies of P. aeruginosa mutants that were increased resistance of persister cells, including muta-
tissues of invertebrates having tested in both C. elegans and G. mellonella have shown tions or alterations in the cellular machinery that is
an open circulatory system. that several mutants showedattenuated pathogenesis in responsible for preprogrammed cell death or apoptosis88.

120 | FEBRUARY 2003 | VOLUME 2 www.nature.com/reviews/drugdisc

 REVIEWS

Evidence, however, points to several factors that could and used to diagnose the condition of a biofilm
be important in such resistance, and continued infection and to direct the administration of an
research is necessary to identify these specific resistance appropriate treatment strategy. As an example, anti-
traits. For instance, biofilm bacteria could present biotic therapy would be more effective if administered
stages when stress-response proteins are produced, during a stage of rapid bacterial growth. Additionally,
conferring increased resistance to environmental con- specific biofilm markers might indicate when it is nec-
ditions (including immune attack, as well as the pres- essary to administer antibiotics prophylactically to
ence of antibiotics). prevent bacteraemia or infection of remote sites. By
The future of treatment strategies for biofilm infec- understanding the physiology of biofilm develop-
tions seems to rest with specifically targeting unique ment, chemotherapeutic agents could be developed to
biofilm characteristics, either in combination with con- promote or prevent transitioning from one stage of
ventional antibiotic therapy or to tailor treatment regi- biofilm maturation to the next by targeting unique
mens to target biofilm infections when they are most biofilm regulatory or signalling molecules. Finally,
susceptible to antibiotics. To develop these novel specific agents might by discovered or developed
methods, research in the future needs to be directed which will interfere with the production of virulence
towards more rigorous studies of the physiological status factors, or promote (or inhibit) the shedding of
of biofilm bacteria during an infection. biofilm bacteria (to coincide with antibiotic therapy)
For instance, biofilms could be found to produce as a particular case requires. Overall, the characteriza-
markers that are associated with the release of shed bac- tion of biofilm development is a crucial component to
teria from the biofilm population. These markers could understanding the diagnosis, treatment and manage-
then be identified with stages in biofilm development ment of biofilm infections.

1. Lawrence, J. R., Korber, D. R., Hoyle, B. D., Costerton, J. W. 15. Brown, M. R., Allison, D. G. & Gilbert, P. Resistance of 29. Jensen, E. T. et al. Complement activation by
& Caldwell, D. E. Optical sectioning of microbial biofilms. bacterial biofilms to antibiotics: a growth-rate related effect? Pseudomonas aeruginosa biofilms. Microb. Pathol. 15,
J. Bacteriol. 173, 6558–6567 (1991). J. Antimicrob. Chemother. 22, 777–783 (1988). 377–388 (1993).
One of the first descriptions of the 3-dimensional 16. Anwar, H., Strap, J. L. & Costerton, J. W. Establishment of 30. Rogers, J., Phillip, F. & Olliff, C. The effects of extracellular
structure of natural biofilms. aging biofilms: possible mechanism of bacterial resistance slime from Staphylococcus epidermidis on phagocytic
2. Lawrence, J. R. & Neu, T. R. Confocal laser scanning to antimicrobial therapy. Antimicrob. Agents Chemother. 36, ingestion and killing. FEMS Immunol. Med. Microbiol. 9,
microscopy for analysis of microbial biofilms. Methods 1347–1351 (1992). 109–115 (1994).
Enzymol. 310, 131–144 (1999). 17. Mah, T. F. & O’Toole, G. A. Mechanisms of biofilm resistance 31. Moran, F. J. et al. Phagocytosis and killing of slime-
3. Lewandowski, Z., Stoodley, P., Altobelli, S. & Fukushima, E. to antimicrobial agents. Trends Microbiol. 9, 34–39 (2001). producing Staphylococcus epidermidis by
Hydrodynamics and kinetics in biofilm systems — recent 18. Sauer, K. & Camper, A. K. Characterization of phenotypic polymorphonuclear leukocytes. Effects of sparfloxacin.
advances and new problems. Proceedings of the Second changes in Pseudomonas putida in response to surface- Rev. Esp. Quimioter. 11, 52–57 (1998).
IAWQ International Specialized Conference on Biofilm associated growth. J. Bacteriol. 183, 6579–6589 (2001). 32. Costerton, J. W., Lewandowski, Z., Caldwell, D., Korber, D.
Reactors. Paris, France, pp. 313–319 (1993). 19. Stewart, P. S. Mechanisms of antibiotic resistance in & Lappin-Scott, H. M. Microbial biofilms. Annu. Rev.
4. DeBeer, D., Stoodley, P., Roe, F. & Lewandowski, Z. bacterial biofilms. Int. J. Med. Microbiol. 292, 107–113 Microbiol. 49, 711–745 (1995).
Effects of biofilm structure on oxygen distribution and mass (2002). 33. Costerton, J. W., Stewart, P. S. & Greenberg, E. P. Bacterial
transport. Biotech. Bioeng. 43, 1131–1138 (1994). An excellent review from an engineering perspective biofilms: a common cause of persistent infections. Science
5. Davies, D. G. et al. The involvement of cell-to-cell signals in on antibiotic resistance in bacterial biofilms. 284, 1318–1322 (1999).
the development of a bacterial biofilm. Science 280, 20. Donlan, R. M. & Costerton, J. W. Biofilms: survival 34. Meluleni, G. J., Grout, M., Evans, D. J. & Pier, G. B. Mucoid
295–298 (1998). mechanisms of clinically relevant microorganisms. Clin. Pseudomonas aeruginosa growing in a biofilm in vitro are
The first paper to link biofilm development with Microbiol. Rev. 15, 167–193 (2002). killed by opsonic antibodies to the mucoid
intercellular communication. A comprehensive review outlining many organisms of exopolysaccharide capsule but not by antibodies produced
6. Martinelli, D., Bachofen, R. & Brandl, H. Effect of medium medical importance and what is known about their during chronic lung infection in cystic fibrosis patients.
composition, flow rate, and signaling compounds on the resistance to specific antibiotics. J. Immuno. 155, 2029–2038 (1995).
formation of soluble extracellular materials by biofilms of 21 Gilbert, P., Maira-Litran, T., McBain, A. J., Rickard, A. H. & 35. Dasgupta, M. K. Biofilm causes decreased production of
Chromobacterium violaceum. Appl. Microbiol. Biotechnol. White, F. W. The physiology and collective recalcitrance of interferon-γ. J. Am. Soc. Nephrol. 7, 877–882 (1996).
59, 278–283 (2002). microbial biofilm communities. Adv. Microb. Physiol. 46, 36. Suci, P. A., Mittelman, M. W., Yu, F. P. & Geesey, G. G.
7. Huber, B. et al. Genetic analysis of functions involved in the 202–256 (2002). Investigation of ciprofloxacin penetration into Pseudomonas
late stages of biofilm development in Burkholderia cepacia A thoughtful and insightful look at the mechanisms aeruginosa biofilms. Antimicrob. Agents Chemother. 38,
H111. Mol. Microbiol. 46, 411–426 (2002). underlying the resistance of biofilms to antibiotic 2125–2133 (1994).
8. National Institute of Dental Research. The Oral Health of treatments. An excellent paper describing a novel in situ method
United States Adults. The National Survey of Dental Caries 22. Gilbert, P., Allison, D. G. & McBain, A. J. Biofilms in vitro and for determining antibiotic penetration into biofilms.
in US Employed Adults and Seniors: 1985–1986. DHHS in vivo: do singular mechanisms imply cross-resistance?. 37. Vrany, J. D., Stewart, P. S. & Suci, P. S. Comparison of
Pub. No. (NIH) 87–2868. (US Department of Health and J. Appl. Microbiol. Suppl. 98, S–110S (2002). recalcitrance to ciprofloxacin and levofloxacin exhibited by
Human Services, Bethesda, Maryland 1987). 23. Yu, F. P. & McFeters, G. A. Rapid in situ assessment of Pseudomonas aeruginosa bofilms displaying rapid-
9. National Institute of Dental Research. Oral health of United physiological activities in bacterial biofilms using fluorescent transport characteristics. Antimicrob. Agents Chemother.
States children. The National Survey of Dental Caries in US probes. J. Microbiol. Methods 20, 1–10 (1994). 41, 1352–1358 (1994).
School Children, 1986–1987. DHHS Pub. No. (NIH) 24. Huang, C.-T., Yu, F. P., McFeters, G. A. & Stewart, P. S. 38. Anderl, J., Franklin, M. J. & Stewart, P. S. Role of antibiotic
89–2247. (Department of Health and Human Services, Nonuniform spatial patterns of respiratory activity within penetration limitation in Klebsiella pneumoniae biofilm
Bethesda, Maryland 1989). biofilms during disinfection. Appl. Environ. Microbiol. 61, resistance to ampicillin and ciprofloxicin. Antimicrob. Agents
10. Passerini, L., Lam, K., Costerton, J. W. & King, E. G. 2252–2256 (1995). Chemother. 44, 1818–1824 (2000).
Biofilms on indwelling vascular catheters. Crit. Care Med. A classic paper that visually demonstrates the effect 39. Stone, G., Wood, P., Dixon, M., Keyhan, M. & Matin, A.
20, 665–673 (1992). of an antimicrobial agent against a biofilm in situ. Tetracycline rapidly reaches all the constituent cells of
11. Pople, I. J., Bayston, R. & Hayward, R. D. Infection of 25. Onarheim, H. et al. Outbreak of multiresistant uropothogenic Escherichia coli biofilms. Antimicrob. Agents
cerebrospinal fluid shunts in infants: a study of etiological Acenetobacter baumannii infection. Tidsskr. Nor. Chemother. 46, 2458–2461 (2002).
factors. J. Neurosurg. 77, 29–36 (1992). Laegeforen. 120, 1028–1033 (2000). 40. Larsen, T. Susceptibility of Porphyromonas gingivalis in
12. Costerton, J. W. et al. Bacterial biofilms in nature and 26. Still, J., Law, E., Friedman, B., Furman, S. & Newton, T. biofilms to amoxicillin, doxycycline and metronidazole. Oral
disease. Annu. Rev. Microbiol. 41, 435–464 (1987). Vancomycin resistant organisms in a burn unit. South Med. J. Microbiol. Immunol. 5, 267–271 (2002).
13. Hoiby, N. et al. in Microbial Biofilms (eds Lappin–Scott, H. M. 94, 810–812 (2001). 41. Campanac, C., Pineau, L., Payard, A., Baziard-Mouysset, G.
& Costerton, J. W.) (Cambridge University Press, 27. Rastegar, L. A., Bahrami, H. H. & Alaghebandan, R. & Roques, C. Interactions between biocide cationic agents
Cambridge, 1995). Pseudomonas infections in Tohid Burn Center, Iran. Burns and bacterial biofilms. Antimicrob. Agents Chemother. 46,
14. Luppens, S. B., Reij, M. W. van der Heijden, R. W., 24, 637–641 (1998). 1469–1474 (2002).
Rombouts, F. M. & Abee, T. Development of a standard test 28. Olson, M. E., Ceri, H., Morck, D. W., Buret, A. G. & 42. Stoodley, P., deBeer, D. & Lappin-Scott, H. M. Influence of
to assess the resistance of Staphylococcus aureus biofilm Read, R. R. Biofilm bacteria: formation and comparative electric fields and pH on biofilm structure as related to the
cells to disinfectants. Appl. Environ. Microbiol. 68, susceptibility to antibiotics. Can. J. Vet. Res. 66, 86–92 bioelectric effect. Antimicrob. Agents Chemother. 41,
4194–4200 (2002). (2002). 1876–1879 (1997).

NATURE REVIEWS | DRUG DISCOVERY VOLUME 2 | FEBRUARY 2003 | 1 2 1

REVIEWS

43. Costerton, J. W., Ellis, B., Lam, K., Johnson, F. & Koury, A. E. An important paper linking phenotypic differences in 78. Grobe, K. J., Zahller, J. J. & Stewart, P. S. Role of dose
Mechanism of electrical enhancement of efficacy of biofilms to antibiotic susceptibility. concentraion in biocide efficacy against Pseudomonas
antibiotics in killing biofilm bacteria. Antimicrob. Agents 60. Schurr, M. J., Martin, D. W., Mudd, M. H. & Deretic, V. Gene aeruginosa. J. Indust. Microbiol. Biotechnol. 29, 10–15
Chemother. 38, 2803–2809 (1994). cluster controlling conversion to alginate-overproducing (2002).
Introduction of a novel strategy using electrical fields phenotype in Pseudomonas aeruginosa: functional analysis 79. Allison, D. G., B. Ruiz, B., SanJose, C., Jaspe, A. & Gilbert, P.
to enhance the effectiveness of antibiotics against in a heterologous host and role in the instability of mucoidy. Extracellular products as mediators of the formation and
biofilms. J. Bacteriol. 176, 3375–3382 (1994). detachment of Pseudomonas fluorescens biofilms. FEMS
44. McLeod, B. R., Fortun, S., Costerton, J. W. & Stewart, P. S. 61. O’Toole, G. A. ,Kaplan, H. B. & Kolter, R. Biofilm formation Microbiol. Lett. 167, 179–184 (1988).
Enhanced bacterial biofilm control using electromagnetic as microbial development. Annu. Rev. Microbiol. 54, 49–79 80. Vats, N. & Lee, S. F. Active detachment of Streptococcus
fields in combination with antibiotics. Methods Enzymol. (2000). mutans cells adhered to epon-hydroxylapatite surfaces
310, 656–670 (1999). 62. Belas, R., Simon, M. & Silverman, M. Regulation of lateral coated with salivary proteins in vitro. Arch. Oral Biol. 45,
45. Sternberg, C. et al. Distribution of bacterial growth activity in flagella gene transcription in Vibrio parahaemolyticus. 305–314 (2000).
flow-chamber biofilms. Appl. Environ. Microbiol. 65, J. Bacteriol. 167, 210–218 (1986). An interesting and significant report on the regulation
4108–4117 (1999). 63. Dagastino, L., Goodman, A. E. & Marshall, K. C. of autodispersion of biofilms.
A landmark paper demonstrating the activity and Physiological responses induced in bacteria adhering to 81. Jackson, D. W. et al. Biofilm formation and dispersion under
positioning of single cells in biofilms in real time. surfaces. Biofouling 4, 113–119 (1991). the influence of the global regulator CsrA of Escherichia coli.
46. Ashby, M. J., Neale, J. E. & Critchley, I. A. Effect of 64. Santos, R., Callow, M. E. & Bott, T. R. The structure of J. Bacteriol. 184, 290–301 (2002).
antibiotics on non-growing planktoinic cells and biofilms of Pseudomonas fluorescens biofilms in contact with flowing 82. Choi, J. Y. et al. Identificationof virulence genes in a
Escherichia coli. J. Antimicrob. Chemotherapy 33, 443–452 systems. Biofouling 4, 319–336 (1991). pathogenic strain of Pseudomonas aeruginosa by
(1994). 65. Davies, D. G., Chakrabarty, A. M. & Geesey, G. G. representational difference analysis. J. Bacteriol. 184,
47. Shigeta, M, G. et al. Permeation of antimicrobial agents Exopolysaccharide production in biofilms: substratum 952–961 (2002).
through Pseudomonas aeruginosa biofilms: a simple activation of alginate gene expression in Pseudomonas 83. Hubank, M. & Schatz, D. G. Identifying differences in mRNA
method. Chemotherapy 43, 340–345 (1997). aeruginosa. Appl. Environ. Microbiol. 59, 1181–1186 (1993). expression by representational difference analysis of cDNA.
48. Shigeta, M., Komatsuzawa, M., Sugai, M., Suginaka, H. & First report of the the detection of specific gene Nucleic Acids Res. 22, 5640–5648 (1994).
Usui, T. Effect of the growth rate of Pseudomonas activity in bacteria at the single-cell level, in real time 84. Jander, G., Rahme, L. G. & Ausubel, F. M. Positive
aeruginosa biofilms on the susceptibility to antimicrobial and in situ. correlation between virulence of Pseudomonas aeruginosa
agents. Chemotherapy 43, 137–141 (1997). 66. Dalton, H. M. & March, P. E. Molecular genetics of bacterial mutants in mice and insects. J. Bacteriol. 182, 3843–3845
49. Tanaka G. et al. Effect of the growth rate of Pseudomonas attachment and biofouling. Biotechnology 9, 252–255 (2000).
aeruginosa biofilms on the susceptibility to antimicrobial (1998). A well-conducted investigation of the differences in
agents: beta-lactams and fluoroquinolones. Chemotherapy 67. Dekkers, L. C. et al. Role of the O-antigen of virulence of an important pathogen in different model
45, 28–36 (1999). lipopolysaccharide, and possible roles of growth rate and of hosts.
50. Spoering A. L. & K. Lewis, K. Biofilms and planktonic cells of NADH:ubiquinone oxidoreductase (nuo) in competitive 85. Tan, M. W., Mahajan-Miklos, S. & Ausubel, F. M. Killing of
Pseudomonas aeruginosa have similar resistance to killing tomato root-tip colonization by Pseudomonas fluorescens Caenorhabditis elegans by Pseudomonas aeruginosa used
by antimicrobials. J. Bacteriol. 183, 6746–6751 (2001). WCS365. Mol. Plant Microbe Interact. 11, 763–771 (1998). to model mammalian bacterial pathogenesis. Proc. Natl
An examination of bacterial resistance based on 68. O’Toole, G. A. & Kolter, R. Flagellar and twitching motility are Acad. Sci. USA 96, 715–720 (1999).
defects in preprogrammed cell death. This paper has necessary for Pseudomonas aeruginosa biofilm 86. Tan, M. W., Rahme, L. G., Sternberg, J. A., Tompkins, R. G.
stimulated considerable debate regarding the development. Mol. Microbiol. 30, 295–304 (1998). & Ausubel, F. M. Pseudomonas aeruginosa killing of
mechanism of biofilm resistance. 69. O’Toole, G. A. & Kolter, R. Initiation of biofilm formation in Caenorhabditis elegans used to identify P. aeruginosa
51. McCoy, W. F. & Costerton, J. W. Fouling biofilm Pseudomonas fluorescens WCS365 proceeds via multiple, virulence factors. Proc. Natl Acad. Sci. USA 96, 2408–2413
development in tubular flow systems. Dev. in Indust. convergent signalling pathways: a genetic analysis. (1999).
Microbiol. 23, 551–558 (1982). Mol. Microbiol. 28, 419–461 (1998). 87. Brooun, A., Liu, S. & Lewis, K. A dose-response study of
52. Morisaki, H. Effect of solid-liquid interface on metabolic 70. DeFlaun, M. F., Oppenheimer, S. R., Streger, S., antibiotic resistance in Pseudomonas aeruginosa biofilms.
activity of E. coli. J. Gen. Appl. Microbiol. 29, 195–204 Condee, C. W. & Fletcher, M. Alterations in adhesion, Antimicrob. Agents Chemother. 44, 640–466 (2000).
(1983). transport, and membrane characteristics in an adhesion- 88. Lewis, K. Riddle of biofilm resistance. Antimicrob. Agents
53. Jeffrey, W. H. & Paul, J. H. Activity of an attached and free- deficient pseudomonad. Appl. Environ. Microbiol. 65, Chemother. 45, 999–1007 (2001).
living Vibrio sp. As measured by thymidine incorporation, 759–765 (1999). 89. Latifi, A., Foglino, M., Tanaka, K., Williams, P. &
p-iodonitrotetrazolium reduction, and ATP/ADP ratios. 71. Genevaux, P., Bauda, P., DuBow, M. S. & Oudega, B. Lazdunski, A. A hierarchical quorum-sensing cascade in
Appl. Environ. Microbiol. 51, 150–156 (1986). Identification of Tn10 insertions in the rfaG, rfaP, and galU Pseudomonas aeruginosa links the transcriptional
54. Davies, D. G. & McFeters, G. A. Growth and comparative genes involved in lipopolysaccharide core biosynthesis that activators LasR and RhlR (VsmR) to expression of the
physiology of Klebsiella oxytoca attached to granular affect Escherichia coli adhesion. Arch. Microbiol. 172, 1–8 stationary-phase sigma factor RpoS. Mol. Microbiol. 21,
activated carbon particles and in liquid media. Microb. Ecol. (1999). 1137–1146 (1996).
15, 165–175 (1988). 72. Pratt, L. A. & Kolter, R. Genetic analysis of Escherichia coli 90. Fuqua, W. C., Winans, S. C. & Greenberg, E. P. Quorum
55. Davies, D. G. & Geesey, G. G. Regulation of the alginate biofilm formation: roles of flagella, motility, chemotaxis and sensing in bacteria: the LuxR-LuxI family of cell density-
biosynthesis gene algC in Pseudomonas aeruginosa during type I pili. Mol. Microbiol. 30, 285–293 (1998). responsive transcriptional regulators. J. Bacteriol. 176,
biofilm development in continuous culture. Appl. Environ. 73. Rodriguez-Herva, J. J., Reniero, D., Galli, E. & Ramos, J. L. 269–275 (1994).
Microbiol. 61, 860–867 (1995). Cell envelope mutants of Pseudomonas putida: 91. May, T. B. et al. Alginate synthesis by Pseudomonas
56. Prigent-Combaret, C., Vidal, O., Dorel, C. & Lejeune, P. physiological characterization and analysis of their ability to aeruginosa: a key pathogenic factor in chronic pulmonary
Abiotic surface sensing and biofilm-dependent regulation of survive in soil. Environ. Microbiol. 1, 479–488 (1999). infections of cystic fibrosis patients. Clin. Microbiol. Rev. 4,
gene expression in E. coli. J. Bacteriol. 181, 5993–6002 74. Steyn B, Oosthuizen, M. C., MacDonald, R., Theron, J. & 191–206 (1991).
(1999). Brozel, V. S. The use of glass wool as an attachment surface 92. Singh, P. et al. Quorum-sensing signals indicate that cystic
57. Whiteley, M. et al. Gene expression in Pseudomonas for studying phenotypic changes in Pseudomonas fibrosis lungs are infected with bacterial biofilms. Nature
aeruginosa biofilms. Nature 413, 860–864 (2001). aeruginosa biofilms by two-dimensional gel electrophoresis. 407, 762–764 (2000).
58. Sauer, K. , Camper, A. K., Ehrlich, G. D., Costerton, J. W. & Proteomics 1, 871–879 (2001).
Davies, D. G. Pseudomonas aeruginosa displays multiple 75. Brözel, V. S., Strydom, G. M. & Cloete, T. E. E. A method for
phenotypes as a biofilm. J. Bacteriol. 184, 1140–1154 the study of de novo protein synthesis in Pseudomonas Online links
(2002). aeruginosa after attachment. Biofouling 8, 195–210 (1995).
A paper describing the different phenotypic stages of 76. Svensater, G., Welin, J., Wilkins, J. C., Beighton, D. & FURTHER INFORMATION
biofilm development using primarily a proteomics Hamilton, I. R. Protein expression by planktonic and biofilm Encyclopedia of Life Sciences: www.els.net
approach. It was the first paper to outline the cells of Streptococcus mutans. FEMS Microbiol. Lett. 205, biofilms
complete life-cycle of a biofilm-forming pathogen. 139–146 (2001). US National Institutes of Health Guide: SBIR/STTR Study and
59. Drenkard, E. & Ausubel, F. M. Pseudomonas biofilm 77. Espinosa-Urgel, M., Salido, A. & Ramos, J. L. Genetic Control of Microbial Biofilms:
formation and antibiotic resistance are linked to phenotypic analysis of functions involved in adhesion of Pseudomonas http://grants2.nih.gov/grants/guide/pa-files/PA-99-084.html
variation. Nature 416, 740–743 (2002). putida to seeds. J. Bacteriol. 182, 2363–2369 (2000). Access to this interactive links box is free online.

122 | FEBRUARY 2003 | VOLUME 2 www.nature.com/reviews/drugdisc