Alginate Lyase Promotes Diffusion of Aminoglycosides through the Extracellular Polysaccharide of Mucoid Pseudomonas aeruginosa

Abstract

We demonstrated that a 2% suspension of Pseudomonas aeruginosa alginate completely blocked the diffusion of gentamicin and tobramycin, but not that of carbenicillin, illustrating how alginate production can help protect P. aeruginosa growing within alginate microcolonies in patients with cystic fibrosis (CF) from the effects of aminoglycosides. This aminoglycoside diffusion barrier was degraded with a semipurified preparation of P. aeruginosa alginate lyase, suggesting that this enzyme deserves consideration as an adjunctive agent for CF patients colonized by mucoid strains of P. aeruginosa.

Pseudomonas aeruginosa is one of the most important opportunistic human pathogens (7). This bacterium is ubiquitous and exhibits innate resistance to a wide range of antimicrobial agents, making infections both common and difficult to treat. Patients with cystic fibrosis (CF) have a multisystem disease due to a biochemical defect in the regulation of epithelial chloride transport (35) which leads to the accumulation of thick mucus in the lungs, causing respiratory congestion and increased susceptibility to bronchopulmonary disease (14, 42). Despite an aggressive host immune response (40), patients with CF usually have chronic pulmonary infections with P. aeruginosa which remain intractable to antibiotic treatment (17, 30, 38), making P. aeruginosa the predominant cause of death in CF patients (41). There is also an increasing awareness of the important role of P. aeruginosa biofilms in the contamination of medical biomaterials, such as catheters and prostheses (6). The persistence of P. aeruginosa in CF patient lungs and the establishment of bacterial biofilms, as well as the bacterial resistance to antibiotic action and host-mediated clearance mechanisms in these microniches, has been attributed to the production of an exopolysaccharide called alginate (6, 17, 22, 29).

Role of alginate.

Alginate produced by mucoid strains of P. aeruginosa is a polyuronic acid exopolysaccharide composed of a linear polymer of β 1-4 linked d-mannuronic and l-guluronic acids with O-acetyl side groups found on C-2 and C-3 of the mannuronic acid residues (the percentage of acetylation varies from strain to strain) (13). Alginate production has been shown to inhibit nonopsonic phagocytosis by monocytes and neutrophils both in vitro (3, 11) and in vivo (2). Alginate also increases bacterial adherence to the respiratory epithelia (8, 31), thereby increasing the rate of colonization within the respiratory tract.

Studies examining the susceptibility and resistance of mucoid P. aeruginosa to antibiotics have yielded conflicting results. Thomassen et al. (37) compared the antibiotic susceptibilities of mucoid and nonmucoid pairs of strains and found that many parental mucoid strains were more sensitive to antibiotics than their nonmucoid revertants. In contrast, Govan and Fyfe (18) reported that mucoid strains were more resistant to antibiotics than their nonmucoid counterparts and suggested that the emergence of the mucoid form of P. aeruginosa in CF patient lungs could be due in part to this increased antibiotic resistance (18). The role of alginate as a barrier to antibiotic penetration is supported by the studies of Bayer et al. (3), who found that alginate production decreases the uptake and early bactericidal effect of aminoglycosides, and by Kumon et al. (20), who reported that alginate blocks the diffusion of positively charged hydrophilic drugs. In contrast, Nichols et al. (25, 26) reported that alginate had only a small role in reducing drug penetration.

Sandwich cup assay.

We decided to directly examine the ability of P. aeruginosa alginate to prevent the diffusion of various antibiotics, using an adaptation of the sandwich cup assay (20). Alginate was isolated from culture supernatants of P. aeruginosa mucoid strains FRD1 and 144M, both originally obtained from the sputum of patients with CF (27, 33) by the protocol described by Franklin and Ohman (12). A log phase culture of P. aeruginosa FRD2 (a spontaneous nonmucoid derivative of FRD1 [28]) grown in nutrient broth (Becton Dickinson & Co., Cockeysville, Md.) at 37°C, adjusted to an optical density at 600 nm of 0.6, was used as the indicator strain. After a culture plate insert with a 0.4-μm-pore-size filter (12 mm in diameter; Millipore Corporation, Bedford, Mass.) was centered in a 100-mm-diameter petri dish, 15 ml of Antibiotic Medium 11 (Difco Laboratories, Detroit, Mich.) at 47°C was mixed with 50 μl of the bacterial suspension and poured into the dish. After the agar had hardened, 200 μl of either heat-solubilized 1% Noble agar (Difco) or purified 144M or FRD1 alginate preparation (at various concentrations) was placed into the insert and allowed to solidify. Finally, 50 μl of various concentrations of either gentamicin, tobramycin, polymyxin B, or carbenicillin (all from Sigma Chemical Co., St. Louis, Mo.) diluted in one-fourth strength Dulbecco’s phosphate-buffered saline (Sigma), pH 8.0, or 50 μl of an antibiotic-alginate lyase mixture was added to the insert. Plates were then incubated overnight at 37°C.

Preliminary experiments determined the optimal levels of antibiotics for use in this sandwich cup assay. Zone sizes representing inhibition of bacterial growth were measured, and the radius (after the contribution of the insert was subtracted) of the zone that was found with Noble agar was defined as representing 100% penetration of antibiotics. The growth inhibition zone size for each antibiotic against strain FRD2 increased linearly as a function of antibiotic concentration (data not shown). Based on these results, we selected antibiotic concentrations which gave consistent growth inhibition radius zone sizes of ∼5.50 mm for all subsequent experiments: tobramycin (100 μg), gentamicin (200 μg), carbenicillin (200 μg), and polymyxin B (800 μg).

Effect of alginate on antibiotic penetration.

We next examined whether alginate can affect the growth inhibitory activity of these antibiotics for strain FRD2. Varying amounts of purified alginate were substituted for Noble agar in the insert cup, and antibiotics were placed on top of the alginate prior to incubation. The antibiotic zone sizes were measured for each alginate concentration and compared to those obtained with 1% Noble agar as a control. Inhibition of antibiotic activity for each alginate concentration was determined as follows: percentage of inhibition = 100 × [1 − (radius with alginate/radius with Noble agar)], and the results are shown in Fig. 1. When used at concentrations below 2%, the alginate prepared from strain 144M demonstrated greater inhibitory activity for tobramycin and for gentamicin than the alginate from FRD1. (We also noted that at lower concentrations, 144M alginate is more viscous than FRD1 alginate.) However, when used at concentrations of ≥2%, both alginates completely blocked the inhibitory activity of both aminoglycosides. Similar results were obtained with 2% alginate and polymyxin B (data not shown). In contrast, neither 2% alginate from strain FRD1 nor 2% alginate from strain 144M blocked the penetration of carbenicillin (data not shown).

FIG. 1.

Effect of alginate on the antibiotic activity of aminoglycosides. The abilities of 200 μg of gentamicin and 100 μg of tobramycin to inhibit the growth of P. aeruginosa FRD2 were examined with either 1% Noble agar (control) or various concentrations of purified P. aeruginosa alginate in the insert cup. Zone sizes determined with 1% Noble agar were considered to be 100% activity, and test zone sizes were compared to these control values. Percentage of inhibition was calculated as follows: 100 × [1 − (radius measured with alginate in the sandwich cup assay/radius with 1% noble agar in the sandwich cup assay)]. Data represent the means ± standard errors of the means based on at least four experiments.

Alginate lyase.

Alginates can be enzymatically depolymerized by lyases (eliminases) which catalyze the cleavage of the glycosidic linkages by β-elimination (32). Alginate lyases are produced by a wide variety of bacteria, including Bacillus circulans (19), Klebsiella pneumoniae (5), various Pseudomonas spp. (36, 39), and P. aeruginosa (4, 9, 21, 24, 34). In previous studies, we have demonstrated that the alginate lyase produced by mucoid strains of P. aeruginosa can degrade alginate (10, 34). To determine whether alginate lyase can influence antibiotic diffusion through alginate, we first obtained a semipurified alginate lyase preparation. We constructed pSM13 by inserting PCR-amplified algL (which encodes the P. aeruginosa FRD1 alginate lyase [34]) into the expression vector pET21c (Novagen, Inc., Madison, Wis.). Escherichia coli BL21(DE3)(pSM13) was then grown in Luria broth-Lennox (Difco) with 100 μg of ampicillin per ml at 25°C overnight, induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and incubated for an additional 6 h at 25°C before being harvested. AlgL was extracted from the bacterial pellets by the heat shock procedure described previously (34). After concentration by ultrafiltration, NaCl was added at a final concentration of 500 mM and the preparation was fractionated on a Sephacryl HR300 column (Pharmacia Biotech, Upsala, Sweden) equilibrated with 25 mM K₂HPO₄-500 mM NaCl (pH 7.0) at 4°C. Column fractions were examined for alginate lyase activity by the thiobarbituric acid assay (34). Enzyme-containing fractions were pooled and dialyzed overnight with 25 mM K₂HPO₄ (pH 7.0) at 4°C, loaded onto a Sepharose HP ion-exchange column (Pharmacia Biotech), and eluted with a continuous NaCl gradient from 0 to 500 mM in 25 mM K₂HPO₄ (pH 7.0) at 4°C. Enzymatically active fractions were pooled, concentrated by ultrafiltration, dialyzed against 25 mM K₂HPO₄ (pH 7.0) at 4°C, and stored at −20°C. Enzyme activity was expressed in enzyme units (EU), where one EU represents the amount of enzyme that produces 1 nmol of β-formyl-pyruvate per min per ml at 37°C (10). Samples were examined by electrophoresis with a 0.75-mm-thick Laemmli 14% sodium dodecyl sulfate-polyacrylamide gel which was silver stained (Fig. 2).

FIG. 2.

Silver-stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis profile of semipurified alginate lyase. Lane 1, molecular weight standards (in thousands); lane 2, crude enzyme preparation; lane 3, enzyme preparation after fractionation on a Sephacryl H300 column; lane 4, enzyme preparation after fractionation on a Sepharose HP ion-exchange column.

Effect of alginate lyase on aminoglycoside diffusion through alginate.

Varying amounts of this alginate lyase preparation were mixed with the indicated aminoglycoside (in a final volume of 50 μl) and layered on top of either 2% FRD1 alginate or 2% 144M alginate in the insert cup. The ability of this enzyme to influence aminoglycoside penetration through the alginate preparations was determined by measuring the growth inhibitory zones produced by the aminoglycosides in the presence or absence of the enzyme. The results obtained with FRD1 alginate are shown in Fig. 3. Increasing amounts of alginate lyase decreased the inhibitory activity of 2% FRD1 alginate for both aminoglycosides. When 10.32 EU of alginate lyase was used, the activity of gentamicin for FRD1 was ∼79% of that seen in the absence of alginate, whereas that of tobramycin was ∼92%. Although a greater number of alginate lyase EU was required to affect aminoglycoside penetration through 144M alginate, 100 AlgL EU restored ∼56% of the growth inhibitory activity of gentamicin for strain FRD2 and ∼84% of the growth inhibitory activity of tobramycin (Fig. 4). These results demonstrate that this enzyme can degrade both alginate preparations and allows these aminoglycosides to penetrate to the P. aeruginosa FRD2 target organisms.

FIG. 3.

Effect of AlgL on alginate inhibition of aminoglycoside activity. The abilities of 200 μg of gentamicin and 100 μg of tobramycin to inhibit the growth of P. aeruginosa FRD2 were examined with either 1% Noble agar (control) or 2% FRD1 alginate with various concentrations of AlgL in the insert cup. Zone sizes measured with 1% noble agar were considered to represent 100% antibiotic activity, and zone sizes obtained with 2% FRD1 alginate plus AlgL were compared to these control values. Data represent the means ± standard errors of the means based on at least three experiments.

FIG. 4.

Effect of AlgL on alginate inhibition of aminoglycoside penetration. The abilities of 200 μg of gentamicin and 100 μg of tobramycin to inhibit the growth of P. aeruginosa FRD2 were examined with either 1% Noble agar (control) or 2% 144 M alginate with various concentrations of AlgL in the insert cup. Zone sizes measured with 1% Noble agar were considered to represent 100% antibiotic activity, and zone sizes obtained with 2% 144 M alginate plus AlgL were compared to these control values. Data represent the means ± standard errors of the means based on at least three experiments.

Discussion.

Our results demonstrate that when used at a final concentration of 2%, alginates prepared from two clinical P. aeruginosa isolates can completely block the ability of gentamicin (at 200 μg), tobramycin (at 100 μg), and polymyxin B (at 800 μg) to inhibit the growth of P. aeruginosa. In contrast, carbenicillin was not inhibited by this concentration of alginate. These results can be interpreted as indicating that the polyanionic alginate acts as an ionic trapping agent for the positively charged aminoglycosides and polymyxin B, a model proposed previously by other investigators (1, 15, 25) and supported by our data.

Gordon et al. (16) found that slime dispersants (such as EDTA or NaCl) can reduce alginate viscosity and facilitate diffusion of the antipseudomonal antibiotics through alginate. In the present study, we demonstrated that alginate lyase, an enzyme purified from P. aeruginosa FRD1, can facilitate diffusion of aminoglycosides to the target bacteria. Although the 144M alginate was somewhat more resistant to the effects of this enzyme than the FRD1 alginate (possibly due to variation in mannuronic-to-guluronic acid ratios or differences in the percentage of acetylation), both alginate preparations were effectively degraded by this enzyme. This result suggests that the blocking activity of the alginate may be due to its ability to form viscous gels as well as to its ionic charge and that when the alginate is degraded, this aminoglycoside-blocking activity is effectively reduced.

We and other investigators have suggested that alginate lyase treatment of alginate-encoated bacteria can render them more sensitive to antibiotics or host defenses (2, 11, 23). The present study addresses this issue with a simple, direct assay, the results of which support the hypothesis that degradation of alginate within the lungs of CF patients may render the infecting pathogens more susceptible to aminoglycoside chemotherapy.