Role of Alginate and Its O Acetylation in Formation of Pseudomonas aeruginosa Microcolonies and Biofilms

Abstract

Attenuated total reflection/Fourier transform-infrared spectrometry (ATR/FT-IR) and scanning confocal laser microscopy (SCLM) were used to study the role of alginate and alginate structure in the attachment and growth of Pseudomonas aeruginosa on surfaces. Developing biofilms of the mucoid (alginate-producing) cystic fibrosis pulmonary isolate FRD1, as well as mucoid and nonmucoid mutant strains, were monitored by ATR/FT-IR for 44 and 88 h as IR absorbance bands in the region of 2,000 to 1,000 cm⁻¹. All strains produced biofilms that absorbed IR radiation near 1,650 cm⁻¹ (amide I), 1,550 cm⁻¹ (amide II), 1,240 cm⁻¹ (P⩵O stretching, C—O—C stretching, and/or amide III vibrations), 1,100 to 1,000 cm⁻¹ (C—OH and P—O stretching) 1,450 cm⁻¹, and 1,400 cm⁻¹. The FRD1 biofilms produced spectra with an increase in relative absorbance at 1,060 cm⁻¹ (C—OH stretching of alginate) and 1,250 cm⁻¹ (C—O stretching of the O-acetyl group in alginate), as compared to biofilms of nonmucoid mutant strains. Dehydration of an 88-h FRD1 biofilm revealed other IR bands that were also found in the spectrum of purified FRD1 alginate. These results provide evidence that alginate was present within the FRD1 biofilms and at greater relative concentrations at depths exceeding 1 μm, the analysis range for the ATR/FT-IR technique. After 88 h, biofilms of the nonmucoid strains produced amide II absorbances that were six to eight times as intense as those of the mucoid FRD1 parent strain. However, the cell densities in biofilms were similar, suggesting that FRD1 formed biofilms with most cells at depths that exceeded the analysis range of the ATR/FT-IR technique. SCLM analysis confirmed this result, demonstrating that nonmucoid strains formed densely packed biofilms that were generally less than 6 μm in depth. In contrast, FRD1 produced microcolonies that were approximately 40 μm in depth. An algJ mutant strain that produced alginate lacking O-acetyl groups gave an amide II signal approximately fivefold weaker than that of FRD1 and produced small microcolonies. After 44 h, the algJ mutant switched to the nonmucoid phenotype and formed uniform biofilms, similar to biofilms produced by the nonmucoid strains. These results demonstrate that alginate, although not required for P. aeruginosa biofilm development, plays a role in the biofilm structure and may act as intercellular material, required for formation of thicker three-dimensional biofilms. The results also demonstrate the importance of alginate O acetylation in P. aeruginosa biofilm architecture.

Many species of bacteria produce extracellular polymers that may facilitate nonspecific adhesion to surfaces and provide the framework for biofilms (7). Alginate is an extracellular polysaccharide produced by a variety of gram-negative bacteria including Azotobacter vinelandii, Pseudomonas fluorescens, and Pseudomonas aeruginosa (11, 19, 22, 23, 44). In chronic pulmonary infections of cystic fibrosis (CF) patients, alginate acts as a virulence factor by encapsulating the P. aeruginosa cells. Alginate provides the bacteria with selective advantages for colonization of the pulmonary tissue, through increased resistance to opsonization and phagocytic engulfment (1, 48, 50) as well as through increased protection from toxic oxygen radicals (31, 51). Alginate likely does not play a role in the specific adhesion of P. aeruginosa to pulmonary tissue. However, it may play a role in formation of the bacterial microcolonies that have been observed in vivo (29).

CF patients are initially colonized with nonmucoid strains of P. aeruginosa that produce little or no alginate. However, over time the majority of isolates from chronic pulmonary infections display a mucoid phenotype, indicative of the hyperproduction of alginate (24, 42). The mechanism for the overexpression of alginate is complex and requires several regulatory proteins that act in a hierarchical regulatory cascade (55). The top of the regulatory hierarchy is mediated by an alternative sigma factor, ς²², encoded by algT (also designated algU), located at 68 min on the P. aeruginosa chromosome (10, 13, 28, 34). P. aeruginosa isolates from sources other than pulmonary tissues usually display the nonmucoid phenotype. The nonmucoid phenotype of these isolates is due to control of ς²² by the anti-sigma factor MucA or MucB (21, 33, 35, 56). The genes for these negative regulators lie on the same operon as algT(U). Pulmonary isolates of P. aeruginosa from CF patients often have mutations in mucA (33). Mutations in this negative regulator of ς²² result in increased expression of algT, by an autoregulatory mechanism (10), and increased expression of the alginate biosynthetic operon (21, 47), which is controlled by the algD promoter (9). Therefore, mutations in mucA result in hyperproduction of P. aeruginosa alginate in CF pulmonary isolates. Upon growth of mucoid CF isolates on laboratory medium, the strains rapidly switch to the nonmucoid phenotype. This switching is often the result of suppressor mutations at the algT locus (10, 46).

The structure of alginate from CF isolates is a linear polymer of d-mannuronic acid (M) and its C5 epimer, guluronic acid (G), linked by β1-4 glycosidic bonds (11, 18). P. aeruginosa alginates are found not as repeating disaccharides but as random blocks of MM residues and MG residues (25, 26). The alginate produced by CF isolates, including P. aeruginosa FRD1, is O acetylated at the C-2 and/or the C-3 positions of the mannuronic acid residues (8, 16, 52). Most of the biosynthetic genes for alginate are located in an operon at 34 min on the P. aeruginosa chromosome (4, 5). The algD gene, which encodes GDP-mannose dehydrogenase, is the first gene in the biosynthetic operon (9). A Tn501 transposon insertion in algD resulted in the nonmucoid phenotype, due to the lack of GDP-mannose dehydrogenase and to the polar effect of the transposon insertion on the downstream alginate biosynthetic genes (4, 45). Genes for the structural modification of alginate also lie on the alginate biosynthetic operon. The products of algI, algJ, and algF are required for the addition of O-acetyl groups to the alginate polymer, and mutations in algI, algJ, or algF resulted in production of an alginate polymer that was not O acetylated (16, 17, 49). Since O acetylation affects the physical properties of alginate, including viscosity, interaction with calcium ions, and the reaction with the mannuronan epimerase and mannuronan lyase (15, 45, 53), alginate O acetylation may affect the ability of P. aeruginosa to form biofilms in vivo.

To provide chemical and structural information on living bacterial biofilms, nondestructive analytical and microscopic methodologies have been developed (40). Attenuated total reflection/Fourier transform-infrared spectroscopy (ATR/FT-IR) is a technique that has provided information about microorganisms and extracellular polymers at solid-liquid interfaces (39, 41). With the ATR/FT-IR technique, an IR spectrometer provides IR radiation that is reflected within an IR-transparent substance (e.g., germanium) termed an internal reflection element (IRE) (Fig. 1). Under certain conditions in which the outside medium has a lower refractive index than the IRE, the reflection generates a field of radiation in the medium outside of the IRE, termed an evanescent field. The intensity of this external evanescent field decays exponentially to zero within approximately 1 μm of the IRE. Molecules of cellular biomass or extracellular polymers within the evanescent field absorb the IR radiation, thereby producing an IR absorbance spectrum. Molecules outside of the evanescent field are not detected (Fig. 1). Therefore, ATR/FT-IR can provide unique spectral information on the innermost portion of biofilms near the liquid-substratum interface without disturbance of the biofilm.

FIG. 1.

Use of ATR/FT-IR for biofilm analysis. An IR beam reflects within an IR-transparent substance (germanium) termed an IRE. The reflection generates a field of radiation (evanescent field) in the medium outside of the IRE. The intensity of the evanescent field decays exponentially to zero within approximately 1 μm of the IRE. Molecules of cellular biomass or extracellular polymers within the evanescent field absorb the IR radiation, thereby producing IR absorption spectra. Molecules outside of the evanescent field are not detected. Experimentally, opposing sides of the IRE are sealed with two stainless steel plates with O rings to create flow channels.

Another technique used to study living biofilm is scanning confocal laser microscopy (SCLM) (30). SCLM allows optical sectioning of biofilms and three-dimensional reconstruction of the SCLM images. The SCLM technique reduces the out-of-focus haze that is produced in light microscopy, resulting from the depth of the focal plain through the biofilm.

In this study, the role of alginate in biofilm development was investigated by on-line examination of P. aeruginosa colonization of surfaces, using ATR/FT-IR and SCLM. Two mucoid and three nonmucoid P. aeruginosa strains were examined here, including the mucoid CF pulmonary isolate FRD1 (43), which has the mucA22 allele and overproduces O-acetylated alginate. The nonmucoid strains examined were derivatives of FRD1 and included an algT18 spontaneous nonmucoid mutant; FRD440, an algT Tn501 mutant; and FRD1131, an algD Tn501 mutant. To determine the role of alginate O acetylation in biofilm formation, a mucoid O-acetylation-deficient algJ mutant was analyzed by ATR/FT-IR and SCLM.

MATERIALS AND METHODS

Bacterial strains, plasmids, and media.

Bacterial strains and plasmids used in this study are shown in Table 1. Escherichia coli and P. aeruginosa were maintained on L agar (10 g of tryptone, 5 g of yeast extract, 5 g of NaCl, and 15 g of Bacto Agar per liter). A 1:1 mixture of Pseudomonas isolation agar (Difco) and L agar was used to select for P. aeruginosa following matings with E. coli. Antibiotics were used at the following concentrations (per milliliter): ampicillin at 100 μg and carbenicillin at 300 μg. The medium for the flow system (1× biofilm medium) contained per liter: 0.09 mM sodium glutamate, 0.5 mM glycerol, 0.02 mM MgSO₄, 0.15 mM NaH₂PO₄, 0.34 mM K₂HPO₄, and 145 mM NaCl. The pH of the medium was adjusted to 7.0. The medium for P. aeruginosa FRD440 also contained histidine at a concentration of 0.02 mM. When bacteria containing plasmid pMF230 were analyzed, 50 μg of carbenicillin per ml was added to the 1× biofilm medium.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Genotype or phenotypea	Source/ reference
E. coli HB101	proA2 leuB6 thi-1 lacY1 hsdR hsdM recA13 supE44 rpsL20	This laboratory
P. aeruginosa
FRD1	Cystic fibrosis isolate, mucA22 (Alg+)	43
FRD2	mucA22 algT18 (Alg−)	14
FRD440	mucA22 hisI algT::Tn501 (Alg−)	14
FRD1131	mucA22 algD::Tn501-33 (Alg−)	45
FRD1153	mucA22 algJ5 (Alg+, lacks O acetylation)	17
Plasmids
pRK2013	ColE1-Tra(RK2)+ Kmr	12
pMF36	Ptrc-expression vector pKK233-2 with oriVSForiT Apr	15
pMF230	pMF36 with mut2 GFP	This study

^aAbbreviations for phenotypes: Alg⁺, alginate overproduction; Alg⁻, nonmucoid; Ap^r, ampicillin resistance; Km^r, kanamycin resistance; and Tra⁺, transfer by conjugation. 

DNA manipulations.

General DNA manipulations were performed as described previously (32). Restriction endonucleases were purchased from Boehringer Mannheim. Triparental matings were used to mobilize plasmids from E. coli to P. aeruginosa, using the conjugative helper plasmid pRK2013 (12). Oligonucleotide primers were synthesized on an Applied Biosystems 380B DNA synthesizer.

To visualize bacteria with SCLM, a plasmid that constitutively expressed the green fluorescent protein (GFP) was introduced into each of the P. aeruginosa strains. The gene for the GFP containing the mut2 mutation (6) was amplified by PCR from plasmid pBCgfp (54). The PCR primers used in the amplification were as follows: GFPSal3, 5′-GCGCGTCGACAGGAGAAGAAAAAATGAGTAAAGGAGAAGA-3′; and GFPHind4, 5′-GTACCTGGAATTCTACGAAGCTTATTTGTATAGTTCATCC-3′. The primers were designed to introduce a P. aeruginosa ribosomal binding site upstream of gfp. The PCR product was digested with SalI and HindIII and ligated into pUC19. The XbaI and HindIII fragment from pUC19, containing the GFP mut2 mutation, was ligated into vector pMF36 (16) behind the trc promoter, forming plasmid pMF230. pMF230 contained an oriT site and the stable replication fragment and therefore could be mobilized into P. aeruginosa by triparental mating, where it was maintained. Since pMF230 does not contain the lacI repressor, gfp was expressed constitutively.

Flow systems.

The flow system for the ATR/FT-IR experiments had three separate flow cells and was designed for use with a multichannel IR spectrometer (41). Specifically, each flow system contained a medium reservoir(s); three pumps, each with two pump heads; silicon tubing including inoculation tubes; three flow cells; and three waste reservoirs. Each flow cell (Harrick Scientific Corp., Ossining, N.Y.) consisted of two flow plates and a germanium (Ge) IRE that was 50 mm by 10 mm by 2 mm, with the entrance and exit windows cut at 45^o angles. Each flow plate contained an O-ring groove with a Viton O ring and a flow-channel groove with an inlet and an outlet (Fig. 1). A flow cell was assembled by mounting flow plates on each side of the IRE, creating sealed flow channels on each side of the IRE. Titanium masks were inserted between the O ring and the IRE, thereby minimizing the O-ring interaction with the evanescent field. The volume of each flow channel was approximately 0.4 ml. The flow cells were assembled, leak tested, and sterilized with ethylene oxide gas. The remainder of the flow system was sterilized by autoclaving.

The flow systems for the SCLM experiments consisted of a medium reservoir, a pump, silicon tubing, a flow cell, and a waste reservoir. The flow cell contained a polycarbonate support with inlet and outlet ports. A glass coverslip was sealed onto the polycarbonate support with a Viton gasket. The biofilms that formed on the glass coverslip were examined by SCLM. Both the ATR/FT-IR and SCLM flow cells were maintained at 37(+1)^oC during biofilm formation.

Inoculation and endpoint analyses.

Prior to injection, each strain was incubated in multiple flasks containing 10× biofilm medium for 16 to 20 h at 37°C. When the cell density reached approximately 10⁷ cells/ml, as determined by absorption/scattering (optical density) at 610 nm and previously determined growth curves, the cultures were diluted with an appropriate amount of sterile 0.85% NaCl solution to make a 10⁶-cells/ml solution. Prior to inoculation, sterile water was pumped through each flow channel to obtain background spectra. After the background spectra stabilized, sterile 1× medium was pumped through the system to condition the IRE. Injection tubing connected to the outlet tubing of each flow cell was filled with 5 ml of an inoculum. The flow cells were inoculated by reversing the direction of flow, to pump the inoculation medium into each flow channel. Once inoculated, the pumps were stopped for 20 min to allow cell attachment. Following inoculation, sterile 1× biofilm medium was pumped into the flow cells at a rate of 1.2 ml/min for 44, 64, or 88 h.

Plating on L agar was used to determine the number of cells on the surfaces at the end of 44-h and certain 88-h experiments. Crystals with attached biofilms were placed in glass dishes, and biofilms were removed with a Teflon policeman. Each solution of the cells was added to a test tube with glass beads, vortexed for 5 min, and plated on L agar. Mucoid colonies were differentiated from nonmucoid colonies by visual inspection.

At the end of a few 88-h experiments, the IRE was carefully removed from the flow cell and air dried in a desiccator. Once dried, the IRE with the attached biofilm was placed into a sample holder and the sample spectrum was determined. The IRE was immediately cleaned and analyzed to collect a background spectrum.

Spectrometer and data processing.

The ATR/FT-IR experiments were performed with a customized Analect FT/IR (Orbital Analect, Pomona, Calif.) spectrometer specifically designed for biofilm monitoring (41). This instrument was composed of three optical channels and used two carousels of mirrors to transfer the IR light to and from each flow cell. The instrument had a mercury-cadmium-telluride detector for increased sensitivity and a refractively scanning interferometer for temperature stability. All raw data were collected at a resolution of 4 cm⁻¹. Each stored interferogram, an average of 256 scans, was Fourier processed with a Mertz apodization function (27). The resultant sample and water reference single-beam spectra were divided by a system background, single-beam spectrum, producing transmission spectra that were then converted to absorption spectra. Water absorbance bands were then interactively subtracted from each spectrum to produce a biofilm spectrum (20). In most cases, the final absorbance spectra were baseline corrected.

SCLM.

Confocal images were collected using a Leica TCS-NT confocal scanning laser microscope equipped with an argon 488-nm laser, 500-nm dichroic beam splitter, and 525/50-nm filter block. This configuration provided optimal signal analysis from the GFP mut2 mutation. An ×10 or ×40 oil immersion lens was used to collect 32 to 64 stacks of images, depending on the thickness of the biofilm. The images were taken with 1,024 by 1,024 bit resolution in the x and y plane. Three-dimensional reconstruction of the images was performed using a maximum projection of the stack.

RESULTS

Spectral analysis of biofilms formed by mucoid, nonmucoid, and O-acetylation-deficient strains of P. aeruginosa.

ATR/FT-IR was used to generate IR spectra from the portion of P. aeruginosa biofilms developing within the approximately 1 μm of germanium surfaces. Figure 2 shows three-dimensional diagrams plotting IR spectra from 2,000 to 1,000 cm⁻¹ (x and y axes) at 4-h intervals (z axis). In all experiments, a solution of approximately 10⁶ cells/ml was inoculated into each flow cell, allowing bacterial contact with the opposing surfaces of a germanium IRE and initial adsorption of the P. aeruginosa cells. The first spectrum in each plot was collected during inoculation. For each strain, no IR bands resulting from the cells in the inoculum were observed. Baseline fluctuations in the region near 1,640 cm⁻¹ are artifacts resulting from subtraction of a water band. After 20 min, sterile medium was pumped through the flow cells to stimulate biofilm development. In control experiments that analyzed sterile medium, only small increases in absorbances near 1,640 (water subtraction artifact), 1,260, and 1,080 cm⁻¹ were detected. A band at 1,260 cm⁻¹ was spurious and possibly resulted from the flow cell O ring. The absorbance at 1,080 cm⁻¹ was typically less than 2 milliabsorbance units (mAU) after 44 h and likely resulted from the adsorption of inorganic phosphate to the germanium crystal (39).

FIG. 2.

Interfacial IR absorption spectra of attachment/growth of the mucoid CF isolate P. aeruginosa FRD1 and alg mutant strains plotted at 4-h intervals. Note scale changes to enhance the spectral features (panel B is fivefold greater than panels A and C). In all experiments, an inoculum of 10⁶ cells/ml was introduced into the flow system at time zero. After 20 min, the sterile medium was pumped though the flow cell to stimulate biofilm development. The fluctuations near 1,640 cm⁻¹ were attributed to the subtraction of a water band. (A) These spectra show the IR bands indicative of P. aeruginosa FRD1 biomass that absorbed near 1,650 cm⁻¹ (amide I), 1,550 cm⁻¹ (amide II), 1,250 cm⁻¹ (P⩵O stretching, C—O—C stretching, and/or amide III vibration), 1,090 cm⁻¹ (P—O and C—OH stretching), 1,060 cm⁻¹ (C—OH stretching of alginate), 1,450 cm⁻¹, and 1,400 cm⁻¹ (due in part to C—H deformations and symmetric stretching of carboxylates, respectively). The bands associated with FRD1 biomass were not detected until 24 h. (B) Interfacial IR absorption spectra of biofilms formed by nonmucoid strain P. aeruginosa FRD440. The IR bands indicative of biomass were detected at 8 h. The bands in the FRD440 spectra were similar to bands from the spectra of the mucoid P. aeruginosa FRD1, except for the absence of the band at 1,060 cm⁻¹. (C) Interfacial attachment/growth profile of mucoid P. aeruginosa FRD1153 algJ, a strain that lacked alginate O acetylation. At approximately 32 h, all of the bands indicative of biomass were detected, including the band at 1,060 cm⁻¹.

The IR spectra of the developing biofilm formed by the mucoid CF pulmonary isolate, FRD1, is shown is Fig. 2A. For the first 24 h, spectra generated during biofilm development contained fluctuations similar to those observed with sterile medium, with the exception of a second component in the 1,100 to 1,000 region that was initially detected at 1,060 cm⁻¹ in the 16-h spectra. After 24 h, IR bands indicative of biomass, as previously defined (36–39), were detected near 1,650 cm⁻¹ (amide I), 1,550 cm⁻¹ (amide II), 1,450 cm⁻¹ (due in part to C—H deformation), 1,400 cm⁻¹ (due in part to symmetric stretch for the carboxylate ion), and 1,250 cm⁻¹ (P⩵O and C—O—C stretching and/or amide III) and in the 1,100- to 1,000-cm⁻¹ region (P—O and C—OH stretching). The lack of these IR bands prior to 24 h indicates that not enough biomass was present in the analysis region to produce a signal. After initial detection, all IR bands associated with the FRD1 biofilms increased proportionally with time, signifying an accumulation of biomass within the analysis region.

Figure 2B shows the three-dimensional diagram plotting the IR spectra of a developing biofilm formed by FRD440, a nonmucoid algT::Tn501 mutant of FRD1. As with the FRD1 cells, no IR bands indicative of biomass were detected in the inoculum. Spectra collected after 8 h produced IR bands indicative of biomass. Compared to mucoid strain FRD1, nonmucoid strain FRD440 produced a biofilm with an interfacial biomass that appeared to be fivefold greater; thus, the scale in Fig. 2B was expanded fivefold. Biofilm spectra tended to increase in intensity with time and remained remarkably similar for the duration of the experiment. Biofilms formed by two other nonmucoid strains, FRD2, a spontaneous algT mutant, and FRD1131, an algD::Tn501 mutant, gave a time course of IR spectra with relative characteristics similar to spectra produced by FRD440 algT::Tn501 biofilms (data not shown); however, spectra of FRD1131 algD::Tn501 biofilms had weaker initial IR bands, indicating that less biomass was present in the analysis region during the initial biofilm development phase (described below).

Figure 2C shows the IR spectra associated with the development of a mucoid P. aeruginosa FRD1153 biofilm. Strain FRD1153 is an algJ mutant strain that does not O acetylate alginate. During the initial 28 h, only the fluctuations consistent with sterile control experiments were observed. At approximately 32 h, IR bands associated with cellular biomass were detected. A second component in the 1,100- to 1,000-cm⁻¹ region was also detected at 1,060 cm⁻¹ in the spectrum obtained at 32 h. As with the other strains, the relative absorbances in each spectrum remained consistent with time. Compared with all other strains tested, the O-acetylation-deficient FRD1153 strain had the weakest absorbance bands, signifying poor attachment and/or surface-associated growth.

Spectral analysis demonstrates alginate within biofilms.

Although spectra from all strains produced similar IR bands, spectra of mucoid strains contained a second band in the 1,100- to 1,000-cm⁻¹ region of the spectrum (P—O and C—OH stretching region) that had much weaker relative absorbance in the spectra of the nonmucoid strains. To further characterize this and other spectral differences between the mucoid strains and the nonmucoid strains, 88-h biofilms of each strain were dehydrated and then analyzed by ATR/FT-IR (Fig. 3). Dehydration collapsed the biofilms, thereby increasing the amount of biomass within the analysis region. As with the hydrated biofilms, the dehydrated FRD1 biofilms produced spectra that contained two bands in the 1,100- to 1,000-cm⁻¹ region (Fig. 3A). However, in the spectrum of the dehydrated FRD1 biofilm, the 1,060-cm⁻¹ band (now absorbing at 1,050 cm⁻¹ due to dehydration) had greater relative absorbance than the 1,080-cm⁻¹ band (shifted from 1,090 cm⁻¹). Furthermore, an additional band at 1,735 cm⁻¹ appeared in the dehydrated FRD1 biofilm that was not readily discerned in the spectrum of the 88-h hydrated FRD1 biofilm. The 1,735-cm⁻¹ band had relatively stronger absorbance than the band found in the spectra of the dehydrated O-acetylation-deficient strain (Fig. 3B) or in those of the nonmucoid strains (Fig. 3C to E). The spectrum for dehydrated FRD1 also showed pronounced increases in absorbance at 1,405 and 1,250 cm⁻¹ relative to the nonmucoid strains. In addition, the 1,650-cm⁻¹ band was broader in the FRD1 spectrum (Fig. 3A) than in the spectra from the other strains, revealing the presence of overlapped IR bands.

FIG. 3.

Spectra of biofilms formed by mucoid and nonmucoid strains of P. aeruginosa following dehydration. Each biofilm was cultured for 88 h and then dried in a desiccator and analyzed. The drying process collapsed the biofilm onto the surface of the IRE. (A) Dehydrated FRD1 biofilm showing IR bands associated with the cells: 1,650 cm⁻¹ (amide I), 1,545 cm⁻¹ (amide II), 1,450 cm⁻¹ (C—H deformations), 1,405 cm⁻¹ (in part due to symmetric stretching of the carboxylate ions), 1,250 cm⁻¹ (P ⩵O stretching, C—O—C stretching, and/or amide III), and the 1,100- to 1,000-cm⁻¹ region (P—O and C—OH stretching). IR bands associated with the alginate were also detected: 1,735 cm⁻¹ (C⩵O stretching of esters), 1,615 cm⁻¹ (asymmetric stretching of the carboxylate ion), 1,405 cm⁻¹ and 1,375 cm⁻¹ (symmetric stretching of the carboxylate ion), 1,250 cm⁻¹ (C—O—C for the ester), and 1,060 cm⁻¹ (C—OH stretching of alcohols). (B) Spectrum of dried mucoid FRD1153 showing the lack of prominent alginate bands. The final three spectra show the IR bands associated with the dried nonmucoid strains FRD2 (C), FRD1131 (D), and FRD440 (E).

To determine if the spectral differences observed in the dehydrated FRD1 biofilms were due to alginate within the biofilm matrix, the spectrum produced following dehydration of FRD2 algT18 was subtracted from spectrum produced by dehydrated FRD1 (Fig. 4A). The subtraction spectrum produced bands at 1,730 cm⁻¹ (C⩵O stretching of esters), 1,615 cm⁻¹ (asymmetric stretching of the carboxylate ion), 1,410 cm⁻¹, 1,375 cm⁻¹ (symmetric stretching of the carboxylate ion), 1,250 cm⁻¹ (C—O—C for the ester), and 1,060 cm⁻¹ (C—OH stretching of alcohols). The same IR bands were present in the IR spectrum of purified FRD1 alginate (Fig. 4B), demonstrating that alginate remained associated with the FRD1 biofilms. However, in the spectra of living hydrated FRD1 biofilms, the alginate bands were not only overlapped with the bands associated with the cells but relatively weaker than those produced with the dehydrated biofilms. This increase in the relative absorbance found after dehydration provides evidence that greater amounts of alginate were associated with the FRD1 cells located at depths greater than the ATR/FT-IR analysis range.

FIG. 4.

(A) Spectral subtraction of dehydrated FRD1 biofilm and FRD2 biofilm. The band observed in the spectrum of the hydrated FRD1 biofilm at 1,060 cm⁻¹ (C—OH stretching) was present in the subtraction spectra. Other IR bands observed in the subtraction spectrum were 1,615cm⁻¹ (asymmetric carboxylate stretching), 1,410 cm⁻¹ (symmetric carboxylate stretching), 1,730 cm⁻¹ (C⩵O stretching), and 1,250 cm⁻¹ (C—O stretching). The IR bands at 1,730 and 1,250 were associated with the O-acetyl groups on the alginate that are linked to the mannuronate backbone by ester bonds. (B) The FT-IR spectrum of purified alginate from P. aeruginosa FRD1, showing bands similar to that observed from the FRD1-FRD2 subtraction spectrum. (C) The FT-IR spectrum of purified alginate from P. aeruginosa FRD1153 algJ. This spectrum was similar to the spectrum of purified FRD1 alginate but lacked IR bands at 1,730 and 1250 cm⁻¹ that resulted from ester bonds of O-acetylated alginate.

Similar experiments were performed on the algJ mutant P. aeruginosa FRD1153 and its purified alginate that lacked O-acetyl groups. The IR spectra of alginate purified from FRD1153 (Fig. 4C) did not have the ester bands at 1,730 and 1,250 cm⁻¹ that were found in the alginate of FRD1. Spectral subtractions of hydrated biofilms of FRD2 from hydrated biofilms FRD1153 produced poor quality spectra (data not shown), resulting from insufficient biomass in the FRD1153 biofilms. Subtraction of dehydrated biofilms also produced poor quality subtracted spectra, suggesting that insufficient quantities of alginate were associated with the 88-h FRD1153 biofilms. This was due to the fact that most of the FRD1153 cells switched to the nonmucoid phenotype by 88 h (see results below).

Spectral subtraction experiments for biofilms composed of the algD mutant FRD1131 and the algT mutants FRD440 and FRD2 produced only noisy spectra, with little or no signal, demonstrating little to no alginate within the biofilms of any of these nonmucoid strains.

Effect of alginate and O acetylation on biofilm growth dynamics.

Biofilm dynamics are dependent on a number of different processes, including attachment of cells from the liquid phase, growth of surface-associated cells, and detachment of the surface-associated cells. FT-IR was used to monitor the dynamics of biofilm formation transpiring within approximately 1 μm of the IRE, the analysis range of the attenuated total reflection sampling technique. The amide II band (1,550 cm⁻¹) was used as a marker for biomass since cellular proteins produce amide II absorbance and since the absorbance from water and alginate did not appreciably affect amide II absorbance. For each strain, the other bands in a spectrum remained generally proportional to amide II with time. With the amide II band, the ATR/FT-IR provided a unique view of biomass fluctuations occurring within the innermost portion of a biofilm. In comparison, endpoint analyses were performed at 44 and 88 h to determine the number of CFU associated with each biofilm for each strain.

For FRD1 biofilms, the amide II band was initially detected after 24 h and continued to increase until the absorbance reached 24 mAU at 88 h (Fig. 5). Between 24 and 64 h, the absorbance of the amide II band increased linearly at the rate of 0.5 mAU/h to 19 mAU (Table 2). Over the next 24 h, the amide II band increased only slightly to a maximum of 23 mAU, indicating that the biomass within the analysis region had somewhat stabilized. From 44 to 88 h, the cell density within the FRD1 biofilm increased 25-fold to 7.7 × 10⁸ CFU/cm², with all of the colonies producing a mucoid phenotype at 44 h and 98% of the cells remaining mucoid at 88 h (Table 2). For comparison, the amide II absorbance increased only fourfold between 44 and 88 h, suggesting that the biofilm dynamics occurring within the analysis region were different from those transpiring within the entire biofilm.

FIG. 5.

Plot of amide II band absorbance (1,550 cm⁻¹) over time for mucoid and nonmucoid strains of P. aeruginosa. Open circles, FRD1; open diamonds, FRD1153 algJ; closed circles, FRD2 algT18; boxes, FRD440 algT:: Tn501–33; open triangles, FRD1131 algD::Tn501–31.

TABLE 2.

CFU and interfacial growth rates

Strain	Genotype	Phenotype	Plate count data (CFU/cm2)a		ATR/FT-IR (amide II absorbance)
			44 h	88 h	Linear rate increased (mAU/h)	Growth ratee [ln (1 + mAU)]/h	Doubling timee (h)
FRD1	mucA22	Mucoid	2.8 (0.9) × 107b	7.7 (0.7) × 108b	0.5 (0.95)	0.06 (0.94)	12
FRD2	mucA22 algT18	Nonmucoid	2.0 × 108	6.6 (1.3) × 108	4.6 (0.98)	0.35 (0.98)	2
FRD440	mucA22 algT::Tn501-33	Nonmucoid	2.6 (1.3) × 108	4.1 (1.0) × 108	5.4 (0.97)	0.35 (0.98)	2
FRD1131	mucA22 algD::Tn501-31	Nonmucoid	1.7 (0.8) × 107	1.0 (0.5) × 109	5.0 (0.99)	0.11 (0.98)	6
FRD1153	mucA22 algJ3	Mucoid (no O-acetylation)	4.6 (1.0) × 106c	6.6 (1.7) × 108c	3.2 (0.99)	0.11 (0.98)	6

^aCFU were determined by scraping the attached cells from the germanium IRE with a Teflon policeman and saline, transferring the cells to a test tube, and vortexing with glass beads for 5 min. The cells were serially diluted in sterile saline and plated on L agar. Standard deviations of three samples are indicated in parentheses. 

^b100% of FRD1 colonies were mucoid at 44 h; 98% of FRD1 colonies were mucoid at 88 h. 

^c100% of FRD1153 colonies were mucoid at 44 h; 35% of FRD1153 colonies were mucoid at 88 h. 

^dLinear rates of increase were calculated from amide II absorbances between 32 and 64 h for FRD1, 8 and 32 h for FRD2 and FRD440, 40 and 52 h for FRD1131, and 60 and 88 h for FRD1153. r² values for linear regressions are indicated in parentheses. 

^eGrowth rates and doubling times were calculated between 24 and 60 h for FRD1, 4 and 16 h for FRD2 and FRD440, 16 and 52 h for FRD1131, and 40 and 88 h for FRD1153. The growth rate refers to the cells within the analysis range of the ATR-FTIR, within 1 μm of the IRE surface. 

With the algT mutant strains, FRD440 and FRD2, the amide II band showed a rapid increase over time. Both strains were initially detected at approximately 8 h and showed a rapid increase until 32 h, after which the absorbance of the amide II for FRD440 and FRD2 stabilized near 150 and 140 mAU, respectively (Table 2). These final absorbance values were approximately sixfold stronger than those for FRD1. The rate of increase of the amide II band was 5.4 mAU/h for FRD440 and 4.6 mAU/h for FRD2, approximately 10 times greater than the rate of increase for the mucoid FRD1 biofilms (Table 2). Interestingly, even though the final absorbances and the rates of increase for the amide II band were different, the total numbers of cells in the biofilms were similar (Table 2). This suggested that most of the FRD1 cells were outside the analysis range of the ATR/FT-IR, as compared to cells of the nonmucoid strains.

In planktonic cultures, using 10× biofilm medium, the generation times of all strains in this study were similar, with doubling times of 2.4 h for FRD1 and 2.0 h for the algJ mutant, FRD1153. For the nonmucoid strains, the doubling times in planktonic culture were 2.0 and 2.8 h for the algT mutants, FRD2 and FRD440, respectively, and 2.0 h for the algD mutant, FRD1131. During biofilm growth, the amide II data were used to calculate growth rates in the biofilms by using a logarithmic plot. Assuming that at some time during the experiment the biomass within the ATR/FT-IR analysis region accumulates at an exponential rate, the biomass accumulation rate can be determined by plotting the natural logarithm of the amide II absorbance versus time during exponential growth. The slope represents the rate at which biomass within the analysis region accumulated exponentially. The biomass accumulation rate within the evanescent field was approximately fivefold less for FRD1 than for the algT mutant strains (Table 2). The biomass doubling time, calculated by dividing the ln (2) by the interfacial growth rate, was 12 h for FRD1, six times greater than the generation times for the algT mutants (Table 2). Since the numbers of cells in the biofilms at 88 h were approximately the same for the mucoid and nonmucoid strains, these results suggest that most of the cell division and accumulation of biomass for the FRD1 cells occurred outside the analysis range of the evanescent field (approximately 1 μm from the surface), as compared to cells of the nonmucoid strains.

The algD mutant, FRD1131, had a sigmoid-shaped response similar to algT mutants but showed an increased lag time in the response of the amide II band (Fig. 5). Following initial detection at 20 h, the amide II band increased at a rate of 5.0 mAU/h and had a maximum amide II absorbance of 157 mAU, which was similar to the results obtained with the algT mutant strains. At 44 h, the number of FRD1131 cells was 20-fold less than the numbers of both algT mutant strains (Table 2), providing evidence that the lag resulted in fewer attached cells. At 88 h, the plate count was 10⁹ CFU/cm², only slightly higher than those for mucoid FRD1 and the algT mutant strains. The growth rate within the analysis region was threefold less than that for the algT mutants, indicating that biofilm dynamics may be different for this strain (Table 2).

The algJ mutant strain, FRD1153, showed the longest delay in the initiation of biofilm development, ranging from 32 to 40 h (Fig. 5). The cell density of FRD1153 was 4.6 × 10⁶ cells/cm² at 44 h (Table 2), less than all other FRD strains. FRD1153 showed a 3.8-mAU/h rate of increase in amide II absorbance after approximately 64 h. The amide II absorbance of FRD1153 ultimately surpassed the absorbance of FRD1 biofilms, indicating that the biomass within the analysis range was greater than that of the parent strain, FRD1. The final cell density of FRD1153 was 6.6 × 10⁸ cells/cm² at 88 h, similar to all other strains tested (Table 2). However, as determined by plate counts and colony morphology analysis, 65% of the cells had switched to the nonmucoid phenotype by 88 h, and therefore FRD1153 phenotypically resembled nonmucoid FRD2 cells more than the original FRD1153 strain.

SCLM shows that alginate is important for P. aeruginosa biofilm structure and that O acetylation is critical for biofilm formation.

To visualize biofilm formation using SCLM, pMF230, encoding constitutive expression of the GFP mut2 mutation (6), was introduced into each of the FRD strains. Figure 6 shows the three-dimensional reconstruction of 48-h biofilms formed by FRD1, FRD2, and FRD1153. The inoculum, medium, and flow rates were similar to those in the ATR/FT-IR experiments.

FIG. 6.

SCLM analysis of 48-h biofilms of P. aeruginosa, as viewed from the top and at an angle. All strains constitutively expressed the mut2 GFP. (A) P. aeruginosa FRD1 viewed at ×100 magnification. (B) P. aeruginosa FRD1 viewed at ×400 magnification. (C) P. aeruginosa FRD2 algT18 viewed at ×400 magnification. (D) P. aeruginosa FRD1153 algJ viewed at ×400 magnification.

The three-dimensional SCLM reconstruction of a biofilm formed by FRD1 is shown in Fig. 6A. The micrograph shows a top and an angled view of the biofilm, examined at ×100 magnification. The micrograph shows nonuniform coverage of the surface by FRD1 microcolonies. An individual FRD1 microcolony is shown in Fig. 6B at ×400 magnification. The top view shows the distribution of bacteria on the surface, with microcolonies covering part of the surface and much of the surface not colonized by bacteria. The angled view shows the depth of an individual microcolony, which was approximately 40 μm in the 48-h biofilms. FRD1 biofilms at 90 h were similar in surface coverage but had depths of approximately 125 μm (data not shown). The micrograph demonstrated that FRD1 biofilms tended to form chains of cells within and between the microcolonies.

The three-dimensional SCLM reconstruction of the FRD2 biofilm at ×400 magnification is shown in Fig. 6C. The top view demonstrates a greater surface coverage by the FRD2 cells compared to the FRD1 biofilms. The cells appear more densely packed. The angled view shows the depth of the biofilm (z axis equals 12 μm). Much of the surface was covered by a monolayer of cells, and some of the surface was covered by shallow microcolonies that were approximately 6 μm in depth. In contrast to the FRD1 biofilms, the FRD2 cells failed to form chains of cells. SCLM images of the other algT mutant, FRD440, and the algD mutant, FRD1131, were similar to those of FRD2, showing shallow densely packed biofilms.

SCLM images of 48-h biofilms formed by the algJ mutant, FRD1153, demonstrated little attachment and/or surface-associated growth of this mutant strain, which was unable to O acetylate alginate. Figure 6D shows a microcolony formed by FRD1153 at ×400 magnification. The microcolonies formed by FRD1153 were less than 16 μm in depth and sparsely covered the surface. The SCLM results were consistent with the ATR/FT-IR results, which also showed little surface-associated growth of this strain.

DISCUSSION

Two nondestructive techniques, ATR/FT-IR and SCLM, were used for the on-line monitoring of the biofilm development of mucoid and nonmucoid strains of P. aeruginosa. The ATR/FT-IR technique was used to obtain chemical information regarding the bacteria directly adjacent to a germanium substratum. SCLM in combination with fluorescent bacteria provided images of the three-dimensional structure of cells within living biofilms. For ATR/FT-IR analysis in an aqueous environment, water was interactively subtracted from each biofilm spectrum, producing IR spectra for living biofilms (20, 39). With attenuated total reflection sampling, the IR radiation is most intense near the interface and rapidly decays (exponentially) with distance from the interface. The analysis range, while difficult to quantify under experimental conditions, can be estimated by calculating the depth of penetration, defined as the distance at which the intensity decreased to 37% of its value at the surface. The depth of penetration depends on the angle of incidence, the refractive indices of the IRE and the outside medium (air or water), and the wavelength of the radiation. For our experiments (i.e., germanium in contact with water), the depth of penetration was calculated to be 0.3 μm for the 1,750-cm⁻¹ band, 0.4 μm for the 1,550-cm⁻¹ band, and 0.6 μm for the 1,050-cm⁻¹ band. The values decrease only slightly when the outside medium is air. For example, the 1550-cm⁻¹ band decreases by only 0.02 μm. With SCLM, images of biofilms could be obtained at distances farther from the surface than could be obtained by ATR/FT-IR sampling. As a result, the two techniques provided complementary information on P. aeruginosa colonization of surfaces and on the role of extracellular polymer in P. aeruginosa surface colonization.

The IR spectra of most cells—whether living or dead, hydrated or dehydrated—have similar overall characteristics. This is not surprising, since the IR spectrum gives information about the IR-active vibrational modes of chemical bonds found in the molecules of cells and since the cells consist of analogous classes of cellular components (proteins, lipids, carbohydrates, nucleic acids, etc.) with similar functional groups. In previous studies, bacterial cellular components as well as living biofilms were analyzed using FT-IR (37) and ATR/FT-IR (38, 39, 41). Proteins have amide vibrational modes that give rise to IR bands near 1,650 cm⁻¹ (amide I), near 1,550 cm⁻¹ (amide II), and in the region of 1,300 to 1,240 cm⁻¹ (amide III). Carbohydrates have C—OH and C—O—C stretch modes that absorb IR radiation in the range of 1,200 to 1,000 cm⁻¹. Carboxylate ions have a C⩵O that absorbs in the range of 1,780 to 1,730 cm⁻¹ and contain C—O stretches that absorb near 1,400 cm⁻¹. Phosphates, found in nucleic acids and phospholipids, have a P⩵O that absorbs at 1,250 cm⁻¹ and overlaps the amide III mode. Phosphates also have a P—O stretching mode that absorbs near 1,080 cm⁻¹ and overlaps the C—OH stretching modes. The C—H bending of CH₃ groups and C—H scissoring of CH₂ groups absorb near 1,460 cm⁻¹. These IR bands, typically found in all cells, were also detected and monitored in P. aeruginosa biofilms described here.

In addition to the IR bands found in all of the P. aeruginosa strains, alginate from FRD1 absorbed IR radiation in six different regions, producing spectra with strong absorbance at 1,730, 1,615, 1,410, 1,375, 1,250, and 1,060 cm⁻¹. The band absorbing at 1,615 cm⁻¹ was assigned to the carboxylate ion asymmetric stretch, the band absorbing at 1,410 cm⁻¹ was assigned to the carboxylate ion symmetrical stretch, and the band absorbing at 1,060 cm⁻¹ was assigned to the C—OH stretch. Each of these bands was associated with the purified alginate polymer from FRD1 and FRD1153. The O-acetyl groups, linked to the mannuronate residues by ester bonds, had a C⩵O stretching at 1,730 cm⁻¹ and a C—O—C stretching at 1,250 cm⁻¹. These IR bands were observed in purified FRD1 alginate, as well as in FRD1 biofilms, but not in purified alginate from the algJ mutant strain, FRD1153.

In the SCLM studies, the cells expressed the GFP. Therefore, it was possible to observe the three-dimensional structure of the biofilms without removing the samples from the flow cells for staining or examination. However, it was possible that the expression of this nonnative protein, GFP, altered the physiology of the bacteria and/or affected the three-dimensional structure of the biofilms. Therefore, in addition to the SCLM studies presented here, we examined the biofilms formed by these strains, but not expressing the GFP, using phase-contrast microscopy. Although it was difficult to obtain focused images of biofilms using phase-contrast microscopy (particularly for the mucoid strains), it was apparent that the non-gfp-expressing strains formed biofilms with three-dimensional structures similar to those observed here. The nonmucoid strains formed shallow biofilms, primarily as monolayers, and the mucoid FRD1 formed microcolonies that extended from the surface into the bulk medium. Strain FRD1153 formed sparse microcolonies similar to the bacteria that expressed the GFP. Therefore, expression of the GFP did not appear to have a significant effect on the P. aeruginosa biofilm formation observed in these SCLM studies.

In this study, the amide II band was used as a marker for interfacial biomass because it was associated with IR absorbance of proteins and, unlike amide I, did not overlap an IR band found in alginate (1,615 cm⁻¹) and water (1,640 cm⁻¹). Furthermore, in a study using Caulobacter species, a linear correlation between the amide II band and direct counts of bacteria was observed in the range of 2 × 10⁵ cells/cm² to 2 × 10⁷ cells/cm² (39). However, once the biofilms had more cells per unit area, the correlation was not observed, probably due to the presence of cells outside the analysis range of the ATR/FT-IR technique. In this study, the amide II band was used to monitor biomass within the evanescent field and to determine a rate of biofilm formation within this field for each of the strains.

Results of both the ATR/FT-IR and SCLM analyses of nonmucoid strains demonstrated that alginate was not required for P. aeruginosa biofilm formation, and therefore alginate did not act as a primary adhesin for the P. aeruginosa cells to these surfaces. FRD2 and FRD440 have nonmucoid phenotypes on agar medium and did not produce detectable levels of alginate when cultured in liquid medium. However, since these mutations were in the regulatory gene, algT, it was possible that FRD2 and FRD440 produced undetectable levels of alginate that may have been required for biofilm formation. For this reason, we assayed biofilm formation by strain FRD1131 that had a Tn501 insertion in the alginate biosynthetic gene algD. FRD1131 was unable to produce alginate due to the inability to produce GDP-mannuronate and to the polar effect of the transposon insertion on the downstream alginate biosynthetic operon. Strain FRD1131 showed a delay in biofilm formation but ultimately formed biofilms with spectra similar to those of FRD2 and FRD440. All three nonmucoid strains had amide II absorbances greater than that of FRD1. The reason for the increased lag time of FRD1131 may have resulted from the nonspecific irreversible adsorption of fewer cells during the initial inoculation phase of the experiments. Once biomass was detected, the rate of increase for the amide II band was similar for all three nonmucoid strains. AlgT is an alternative sigma factor and therefore may be responsible for regulation of cellular functions other than alginate production. The ATR/FT-IR results presented here demonstrated that AlgT did not regulate cellular functions that were required for P. aeruginosa biofilm formation, since the algT mutants, FRD2 and FRD440, formed biofilms with amide II absorbances greater than that of FRD1. The SCLM studies confirmed the ability of the algT and algD mutant strains to form biofilms. Two conclusions can be drawn from analysis of nonmucoid strains: (i) alginate was not required for interfacial adhesion/growth; and (ii) the regulatory protein AlgT was not required for biofilm development.

Although alginate production did not affect the ability of P. aeruginosa to attach to or grow on surfaces, alginate significantly affected the architecture of the P. aeruginosa biofilms. The ATR/FT-IR results demonstrated that the nonmucoid strains had amide II absorbances that were as much as eightfold greater than for the mucoid FRD1 strain. However, cell counts revealed approximately the same numbers of cells per unit surface area for FRD1 and for the nonmucoid strains after 88 h. Since the ATR/FT-IR detected only IR bands within approximately 1 μm from the surface, the results suggested that most of the FRD1 cells extended farther from the surface than those of the nonmucoid strains. The SCLM study confirmed those results. The nonmucoid strain FRD2 formed fairly uniform biofilms that completely covered the surface after 48 h of biofilm growth. However, the mucoid FRD1 cells had less surface coverage. The FRD1 biofilms extended farther from the surface and into the medium bulk phase, resulting in approximately the same number of bacteria per unit surface area, but with greater volume and less effective surface coverage. The SCLM studies demonstrated that the nonmucoid P. aeruginosa cells were packed more densely than the mucoid FRD1 cells. This may have been due to the presence of alginate in the space between the individual cells. The ATR/FT-IR subtraction spectra revealed that the alginate produced by the FRD1 cells remained associated with the microcolonies, but primarily outside the analysis range in hydrated biofilms. Therefore, alginate may act as the intercellular matrix that enables the P. aeruginosa cells to extend from the surface and form microcolonies.

Although alginate was not required for biofilm formation, a mutant form of alginate, lacking O-acetyl groups, apparently inhibited P. aeruginosa attachment to or growth on the surfaces used here. During planktonic growth, the algJ mutant strain, FRD1153, had growth rates similar to those of FRD1. However, FRD1153 showed less surface-associated attachment and/or growth. Evidence for the inhibition of biofilm growth by the mutant form of alginate included (i) the ATR/FT-IR studies, which showed less interfacial absorbance of the amide II band, as well as the other IR bands associated with P. aeruginosa cells; (ii) viable cell counts, which demonstrated fewer FRD1153 cells associated with the surface while the cells remained mucoid; and (iii) the SCLM studies, which demonstrated that the FRD1153 cells sparsely populated the surface after 48 h of growth. FRD1153 showed significant surface-associated growth only after a portion of the cells switched to the nonmucoid phenotype. The mechanism for inhibition of biofilm formation on these surfaces by the mutant form of alginate is not known. However, the chemical and physical properties of non-O-acetylated alginate are very different from those of the O-acetylated form. Differences include the reduced viscosity of the deacetylated alginate (53) and increased susceptibility of the deacetylated alginate to degradation by alginate lyases (2, 3, 45). These properties of the mutant alginate may have provided a less stable intercellular matrix for microcolony formation than the O-acetylated alginate.

P. aeruginosa microcolonies have been observed in vivo associated with pulmonary tissue isolated from CF patients (29). The experiments reported here demonstrated that alginate plays a role in the formation of three-dimensional microcolonies in vitro and suggest that alginate may play a similar role in vivo. Since alginate appears to be associated with the biofilm matrix, this polymer may exert its antiphagocytic effect on microcolonies, rather than on individual P. aeruginosa cells.