Chromosomal beta-lactamase is packaged into membrane vesicles and secreted from Pseudomonas aeruginosa.

Membrane vesicles were isolated from one beta-lactam-sensitive and three beta-lactam-resistant Pseudomonas aeruginosa clinical isolates from patients with cystic fibrosis. The presence of the chromosomally encoded beta-lactamase in the membrane vesicles was shown by electron microscopy and enzymatic studies. This is the first report of extracellular secretion of beta-lactamase in P. aeruginosa and it seems that the enzyme is packaged into membrane vesicles.

Delivery of the non-membrane-permeative antibiotic gentamicin into mammalian cells by using Shigella flexneri membrane vesicles.

We developed a model to test whether non-membrane-permeative therapeutic agents such as gentamicin could be delivered into mammalian cells by means of bacterial membrane vesicles. Many gram-negative bacteria bleb off membrane vesicles (MVs) during normal growth, and the quantity of these vesicles can be increased by brief exposure to gentamicin (J. L. Kadurugamuwa and T. J. Beveridge, J. Bacteriol. 177:3998-4008, 1995), which can be entrapped within the MVs. Gentamicin-induced MVs (g-MVs) were isolated from Shigella flexneri and contained 85 +/- 2 ng of gentamicin per microgram of MV protein. Immunogold electron microscopic labeling of thin sections with antibodies specific to S. flexneri lipopolysaccharide (LPS) demonstrated the adherence and subsequent engulfment of MVs by the human Henle 407 intestinal epithelial cell line. Further incubation of g-MVs with S. flexneri-infected Henle cells revealed that the g-MVs penetrated throughout the infected cells and reduced the intracellular pathogen by approximately 1.5 log10 CFU in the first hour of incubation. Antibiotic was detected in the cytoplasms of host cells, indicating the intracellular placement of the drug following the penetration of g-MVs. Soluble antibiotic, added as a fluid to the tissue culture growth medium, had no effect on intracellular bacterial growth, confirming the impermeability of the cell membranes of the tissue to gentamicin. Western blot analysis of MVs with S. flexneri Ipa-specific antibodies demonstrated that the invasion protein antigens IpaB, IpaC, and IpaD were present in MVs. Being bilayered, with outer faces composed of LPS and Ipa proteins, these MVs were readily engulfed by the otherwise impermeable membranes and eventually liberated their contents into the cytoplasmic substance of the host tissue.

S-layered Aneurinibacillus and Bacillus spp. are susceptible to the lytic action of Pseudomonas aeruginosa membrane vesicles.

When S-layered strains of Bacillus stearothermophilus and Aneurinibacillus thermoaerophilus, possessing S-layers of different lattice type and lattice constant as well as S-(glyco)protein chemistry, and isogenic S-layerless variants were subjected to membrane vesicles (MVs) from P. aeruginosa during plaque assays on plates or CFU measurements on cell suspensions, all bacterial types lysed. Electron microscopy of negative stains, thin sections, and immunogold-labelled MV preparations revealed that the vesicles adhered to all bacterial surfaces, broke open, and digested the underlying peptidoglycan-containing cell wall of all cell types. Reassembled S-layer did not appear to be affected by MVs, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis showed that the S-(glyco)proteins remained intact. meso-Diaminopimelic acid, as a peptidoglycan breakdown product, was found in all culture supernatants after MV attack These results suggest that even though MVs are much larger than the channels which penetrate these proteinaceous arrays, S-layers on gram-positive bacteria do not form a defensive barrier against the lytic action of MVs. The primary mode of attack is by the liberation from the MVs of a peptidoglycan hydrolase, which penetrates through the S-layer to digest the underlying peptidoglycan-containing cell wall. The S-layer is not affected by MV protease.

PlcR1 and PlcR2 are putative calcium-binding proteins required for secretion of the hemolytic phospholipase C of Pseudomonas aeruginosa.

The plcHR operon of Pseudomonas aeruginosa includes the structural gene for the hemolytic phospholipase C,plcH (previously known as plcS), and two overlapping, in-phase, genes designated plcR1 and plcR2. Hemolytic and phospholipase C (PLC) activities produced by Escherichia coli and P. aeruginosa T7 expression systems were measured in strains carrying both plcH and plcR genes and in strains carrying each gene separately. When plcH was expressed by itself in the E. coli T7 system, the area of the hemolytic zone on blood agar was less than twice the area of growth. By contrast, when plcR was coexpressed with plcH in this system, the area of the hemolytic zone was approximately 10 times that of the area of the growth on blood agar. Native polyacrylamide gel electrophoretic analyses of PlcH activity expressed in either the E. coli or the P. aeruginosa T7 system carrying plcH alone, or along with the plcR genes, suggest that PlcR either posttranslationally alters the physical or biochemical nature of PlcH or releases PlcH from a complex in the cell so that it can be secreted. The hypothesis that PlcR is involved in the secretion of PlcH is supported by the observation that the ratio of extracellular to cell-associated PlcH activity produced by P. aeruginosa strains containing an in-frame deletion in the chromosomal plcR genes is significantly reduced in comparison with this ratio seen with the wild-type parental strain. This defect in the secretion of PlcH can be complemented by the plcR genes in trans. Additional data suggest that PlcR does not directly affect the secretion of the nonhemolytic phospholipase C (PlcN). PlcR is highly similar to a calcium-binding protein (CAB) from Streptomyces erythraeus. PlcR and CAB contain typical motifs (EF hands) characteristic of eucaryotic calcium-binding proteins, including calmodulin. P. aeruginosa naturally produces membrane vesicles (MVs) containing extracellular proteins including PLC. MVs from the PAO1WT strain contained at least 10-fold more PLC specific activity than those isolated from a strain carrying a deletion of plcR (PAO1 deltaR). Immunogold electron microscopy of PAO1WT and PAO1 deltaR whole cells revealed a distribution of PlcH in these strains consistent with the hypothesis that PlcR is required for the secretion of PlcH.

Interactions between biofilms and the environment.

The surfaces of bacteria are highly interactive with their environment. Whether the bacterium is Gram-negative or Gram-positive, most surfaces are charged at neutral pH because of the ionization of the reactive chemical groups which stud them. Since prokaryotes have a high surface area-to-volume ratio, this can have surprising ramifications. For example, many bacteria can concentrate dilute environmental metals on their surfaces and initiate the development of fine-grained minerals. In natural environments, it is not unusual to find such bacteria closely associated with the minerals which they have helped develop. Bacteria can be free-living (planktonic), but in most natural ecosystems they prefer to grow on interfaces as biofilms; supposedly to take advantage of the nutrient concentrative effect of the interface, although there must also be gained some protective value against predators and toxic agents. Using a Pseudomonas aeruginosa model system, we have determined that lipopolysaccharide is important in the initial attachment of this Gram-negative bacterium to interfaces and that this surface moiety subtly changes during biofilm formation. Using this same model system, we have also discovered that there is a natural tendency for Gram-negative bacteria to concentrate and package periplasmic components into membrane vesicles which bleb-off the surface. Since some of these components (e.g., peptidoglycan hydrolases) can degrade other surrounding cells, the vesicles could be predatory; i.e., a natural system by which neighboring bacteria are targeted and lysed, thereby liberating additional nutrients to the microbial community. This obviously would be of benefit to vesicle-producing bacteria living in biofilms containing mixed microbial populations.

Natural release of virulence factors in membrane vesicles by Pseudomonas aeruginosa and the effect of aminoglycoside antibiotics on their release.

Pseudomonas aeruginosa (and various other gram-negative pathogens) liberate membrane vesicles during normal growth. These bilayered vesicles consist of endotoxin (lipopolysaccharide), outer membrane proteins and several potent hydrolytic enzymes including protease, alkaline phosphatase, phospholipase C and peptidoglycan hydrolase. The vesicles contain pro-elastase and alkaline phosphatase (which are periplasmic constituents) and so are important for packaging periplasmic components as they are liberated to the outside of the cell. Once liberated, the vesicles are capable of fusing with the membranes of epithelial cells and liberating their virulence factors into host cells where they degrade cellular components, thereby aiding infection by the pathogen. The aminoglycoside antibiotic, gentamicin, is thought to kill bacteria by inhibiting protein synthesis, yet this cationic antibiotic can also perturb the packing order of lipids, thereby destabilizing bilayered membranes. For pathogens with highly anionic lipopolysaccharide on their surface, such as P. aeruginosa, this membrane destabilization can be so serious that it can cause cell lysis; these cells are therefore killed by a combination of protein synthesis inhibition and surface perturbation. By destabilizing the membranes of P. aeruginosa, gentamicin increases the release of membrane vesicles three- to five-fold. This may help account for some of the bacterium-mediated toxicity encountered during patient treatment with aminoglycoside antibiotics.

Bacteriolytic effect of membrane vesicles from Pseudomonas aeruginosa on other bacteria including pathogens: conceptually new antibiotics.

Pseudomonas aeruginosa releases membrane vesicles (MVs) filled with periplasmic components during normal growth, and the quantity of these vesicles can be increased by brief exposure to gentamicin. Natural and gentamicin-induced membrane vesicles (n-MVs and g-MVs, respectively) are subtly different from one another, but both contain several important virulence factors, including hydrolytic enzyme factors (J. L. Kadurugamuwa and T. J. Beveridge, J. Bacteriol. 177:3998-4008, 1995). Peptidoglycan hydrolases (autolysins) were detected in both MV types, especially a periplasmic 26-kDa autolysin whose expression has been related to growth phase (Z. Li, A. J. Clarke, and T. J. Beveridge, J. Bacteriol. 178:2479-2488, 1996). g-MVs possessed slightly higher autolysin activity and, at the same time, small quantities of gentamicin. Both MV types hydrolyzed isolated gram-positive and gram-negative murein sacculi and were also capable of hydrolyzing several glycyl peptides. Because the MVs were bilayered, they readily fused with the outer membrane of gram-negative bacteria. They also adhered to the cell wall of gram-positive bacteria. g-MVs were more effective in lysing other bacteria because, in addition to the autolysins, they also contained small amounts of gentamicin. The bactericidal activity was 2.5 times the MIC of gentamicin, which demonstrates the synergistic effect of the antibiotic with the autolysins. n-MVs were capable of killing cultures of P. aeruginosa with permeability resistance against gentamicin, indicating that the fusion of n-MV to the outer membrane liberated autolysins into the periplasm, where they degraded the peptidoglycan and lysed the cells. g-MVs had even greater killing power since they liberated both gentamicin and autolysins into these resistant cells. These findings may help develop a conceptually new group of antibiotics designed to be effective against hard-to-kill bacteria.

Periplasm, periplasmic spaces, and their relation to bacterial wall structure: novel secretion of selected periplasmic proteins from Pseudomonas aeruginosa.

A brief overview of thin sections of cryopreserved walls from select eubacteria will be presented to suggest that all bacteria have functional periplasms, but that these are not necessarily confined to a periplasmic space such as found in typical gram-negative bacteria. Pseudomonas aeruginosa contains many components in its periplasmic space, some of which are required for infection. Throughout its growth cycle, P. aeruginosa blebs-off membrane vesicles that can possess DNA, endotoxin, phospholipase, protease, hemolysin, alkaline phosphatase, and autolysin, each of which must have a molecular phase that resides in the periplasm. These membrane packets make good delivery systems to convey these components to other bacteria and, possibly, tissue. Aminoglycoside antibiotics, such as gentamicin, produce a serious perturbation on the bacterium's surface (separate from the ribosomal effect), which contributes to the killing of the microorganism. Antibiotics such as this increase the size and number of the membrane blebs, which could contribute to septic shock of patients under drug therapy.

Virulence factors are released from Pseudomonas aeruginosa in association with membrane vesicles during normal growth and exposure to gentamicin: a novel mechanism of enzyme secretion.

Pseudomonas aeruginosa blebs-off membrane vesicles (MVs) into culture medium during normal growth. Release of these vesicles increased approximately threefold after exposure of the organism to four times the MIC of gentamicin. Natural and gentamicin-induced membrane vesicles (n-MVs and g-MVs and g-MVs, respectively) were isolated by filtration and differential centrifugation, and several of their biological activities were characterized. Electron microscopy of both n-MVs and g-MVs revealed that they were spherical bilayer MVs with a diameter of 50 to 150 nm. Immunoelectron microscopy and Western blot (immunoblot) analysis of the vesicles demonstrated the presence of B-band lipopolysaccharide (LPS), with a slightly higher proportion of B-band LPS in g-MVs than in n-MVs. A-band LPS was occasionally detected in g-MVs but not in n-MVs. In addition to LPS, several enzymes, such as phospholipase C, protease, hemolysin, and alkaline phosphatase, which are known to contribute to the pathogenicity of Pseudomonas infections were found to be present in both vesicle types. Both types of vesicles contained DNA, with a significantly higher content in g-MVs. These vesicles could thus play an important role in genetic transformation and disease by serving as a transport vehicle for DNA and virulence factors and are presumably involved in septic shock.

Surface action of gentamicin on Pseudomonas aeruginosa.

The mode of action of gentamicin has traditionally been considered to be at the 30S ribosomal level. However, the inhibition of bacterial protein synthesis alone appears to be insufficient to entirely explain the bactericidal effects. Bacteriolysis is also mediated through perturbation of the cell surface by gentamicin (J.L. Kadurugamuwa, J.S. Lam, and T.J. Beveridge, Antimicrob. Agents Chemother. 37:715-721, 1993). In order to separate the surface effect from protein synthesis in Pseudomonas aeruginosa PAO1, we chemically conjugated bovine serum albumin (BSA) to gentamicin, making the antibiotic too large to penetrate through the cell envelope to interact with the ribosomes of the cytoplasm. Furthermore, this BSA-gentamicin conjugate was also used to coat colloidal gold particles as a probe for electron microscopy to study the surface effect during antibiotic exposure. High-performance liquid chromatography confirmed the conjugation of the protein to the antibiotic. The conjugated gentamicin and BSA retained bactericidal activity and inhibited protein synthesis on isolated ribosomes in vitro but not on intact cells in vivo because of its exclusion from the cytoplasm. When reacted against the bacteria, numerous gentamicin-BSA-gold particles were clearly seen on the cell surfaces of whole mounts and thin sections of cells, while the cytoplasm was devoid of such particles. Disruption of the cell envelope was also observed since gentamicin-BSA and gentamicin-BSA-gold destabilized the outer membrane, evolved outer membrane blebs and vesicles, and formed holes in the cell surface. The morphological evidence suggests that the initial binding of the antibiotic disrupts the packing order of lipopolysaccharide of the outer membrane, which ultimately forms holes in the cell envelope and can lead to cell lysis. It is apparent that gentamicin has two potentially lethal effects on gram-negative cells, that resulting from inhibition of protein synthesis and that resulting from surface perturbation; the two effects in concert make aminoglycoside drugs particularly effective antibiotics.

Interaction of gentamicin with the A band and B band lipopolysaccharides of Pseudomonas aeruginosa and its possible lethal effect.

The lipopolysaccharide (LPS) of Pseudomonas aeruginosa PAO1 possesses two distinct types of O polysaccharide, A and B band LPSs, but the majority of clinical isolates from cystic fibrosis patients who are infected with the organism possess only the A band as the major LPS antigen. The initial step in a series of events during the uptake of aminoglycoside antibiotics such as gentamicin is the ionic binding of the molecule to the cell surface. In an attempt to elucidate the role of A and B band LPSs of P. aeruginosa in this passive ionic binding of gentamicin to the outer membrane and its possible lethal effects, strains PAO1 (A+B+) and LPS isogenic derivatives (A+B-,A-B+,A-B-) were treated with the antibiotic. Ionic binding of gentamicin appeared to be subtly different in PAO1 and its LPS derivatives; a lethal dose of drug was bound to all strains, although the degree of binding varied with each strain. The outer membrane affinity for gentamicin was higher in strains possessing the B band than in strains with A band LPS, and these B band strains were more prone to antibiotic-induced killing. Strains with both A and B band LPSs bound the most gentamicin of all strains, and this binding caused an almost 50% loss in viability. Ionic binding of aminoglycoside antibiotucs to the outer membrane of cell surfaces must not only weaken th cell surface (R. E. W. Hancock, Annu. Rev. Microbiol. 38:237-264, 1984; N. L. Martin and T. J. Beveridge, Antimicrob. Agents Chemother. 29:1079-1087, 1986; S. G. Walker and T. J. Beveridge, Can. J. Microbiol. 34:12-18, 1988) but it must also be more important in cell death than was originally thought.

Antibody responses in the lungs of mice following oral immunization with Salmonella typhimurium aroA and invasive Escherichia coli strains expressing the filamentous hemagglutinin of Bordetella pertussis.

The filamentous hemagglutinin (FHA) of Bordetella pertussis was expressed in the attenuated aroA mutant of Salmonella typhimurium, SL3261, and in a strain of Escherichia coli harboring Shigella flexneri plasmid pWR110, which encodes bacterial invasiveness for epithelial cells. Expression of FHA in these strains did not interfere with their ability to invade Henle cells. Immunoglobulins A and G specific for FHA were detected in lung washes of mice following oral immunization with the live recombinant organisms; antibody levels were significantly higher than those in mice immunized with killed bacteria administered orally or intraperitoneally. Live oral vaccines carrying protective antigens of B. pertussis may be an important alternative to new-generation component vaccines against whooping cough.

Intercellular spread of Shigella flexneri through a monolayer mediated by membranous protrusions and associated with reorganization of the cytoskeletal protein vinculin.

The spread of Shigella flexneri in a monolayer of infected Henle and HeLa cells was studied by using immunofluorescence and electron microscopy. Infected cells produced numerous bacterium-containing membranous protrusions up to 18 microns in length that penetrated adjacent cells and were subsequently phagocytosed. Fluorescence staining of actin and vinculin in infected cells with phalloidin and monoclonal antibody to vinculin, respectively, demonstrated that the protrusions containing the bacteria consisted of these cytoskeletal proteins. Actin accumulated predominantly at the poles of bacteria distal to the tip of protrusions and appeared as trails extending back towards the host cell cytoplasm. Vinculin, however, was distributed uniformly around the bacteria and throughout the protrusion. A profound rearrangement of vinculin occurred in Henle and HeLa cells following infection with shigellae: whereas in uninfected cells it was distributed mainly around the cell periphery, in infected cells it concentrated mainly around clusters of bacteria in the cytoplasm. This suggests a possible involvement of the vinculin cytoskeletal protein in the intercellular spread of shigellae during an infection.

Inhibition of complement-factor-5a-induced inflammatory reactions by prostaglandin E2 in experimental meningitis.

The possible role of complement factor 5a (C5a) and prostaglandin E2 (PGE2) in cerebrospinal fluid (CSF) pleocytosis and protein accumulation was assessed in a rabbit model of meningitis. Intracisternally administered C5a caused a rapid, early influx of leukocytes into CSF that peaked at 1 h after injection; by 6 h, cell counts were slightly higher than those in controls. Administration of PGE2 or saline did not induce detectable CSF leukocytosis. Coadministration of PGE2 with C5a decreased CSF leukocytosis in a dose-related fashion. Protein concentration increased 30 min after administration of C5a, peaked after 1 h, and remained elevated for 6 h. PGE2 caused a dose-related increase in protein content after 2 h, whereas coadministration caused an inversely dose-related inhibition of the C5a-induced protein influx into CSF. These data suggest that PGE2 in the subarachnoid space exerts an inhibitory action on the C5a-mediated response that is probably not related to its direct effects on protein extravasation.

Cerebrospinal fluid protein profile in experimental pneumococcal meningitis and its alteration by ampicillin and anti-inflammatory agents.

We studied the effects of ampicillin and anti-inflammatory agents on cerebrospinal fluid (CSF) protein patterns in rabbits with meningitis caused by Streptococcus pneumoniae. CSF proteins were analyzed in the acute phase of infection and during convalescence by using sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by two types of staining or by immunoblotting. During the acute phase a massive influx of serum albumin into the CSF was accompanied by the appearance of other proteins of high and low molecular weight. In untreated animals and animals given ampicillin alone or together with indomethacin, the abnormal pattern persisted for up to 30 d. Combined treatment with ampicillin and dexamethasone or oxindanac, a new nonsteroidal anti-inflammatory drug, partially prevented the alterations in CSF protein patterns. Reversion to normal CSF protein patterns was fastest in these two groups.

Influence of cephalosporins and iron on surface protein antigens of Klebsiella pneumoniae in vivo.

The outer membrane protein (OMP) profiles of Klebsiella pneumoniae grown in a rabbit peritonitis model in the presence or absence of cephalosporins were investigated. Six high-molecular-weight OMPs (Mr 69,000 to 83,000) were induced under iron-depleted conditions in vitro. Three of these proteins (the 69,000-Mr protein [69K protein] and the 70K and 78K proteins) and trace amounts of the 73K and 75K proteins were induced in the OM of bacteria infecting the peritoneal cavity of rabbits. Addition of iron either to the growth medium in vitro or to the peritoneum in vivo repressed the expression of these proteins. Cephaloridine had no significant effect on the OMP profiles. An additional 56,000-Mr protein was observed in the OM of bacteria cultivated in vivo in the presence of CGP 17520 and also to a lesser extent in vivo under conditions of iron excess. A difference in recognition of OM antigens between cells grown in vitro and in vivo was observed by immunoblotting techniques. The 26K, 27.5K, and 28.5K antigens present in the OM of cells grown in vitro (but not in vivo) were recognized by antibodies raised against bacteria cultivated in vitro under conditions of iron depletion, but were not recognized by antisera raised against bacteria harvested directly from infections. Antisera raised against a nonencapsulated K. pneumoniae strain caused no agglutination of encapsulated K. pneumoniae grown in vivo in the absence of cephalosporins. Rapid agglutination was observed with this antiserum when the same encapsulated strain was grown in vivo in the presence of either cephalosporin, indicating less occlusion of critical antigens by the capsule.

Media for study of growth kinetics and envelope properties of iron-deprived bacteria.

Ion-exchange chromatography was used to remove iron from complex and chemically defined laboratory media. The kinetics of metal cation removal from the media was investigated by using atomic absorption spectrophotometry, and the results indicated that over 90% of the iron could be eliminated from certain complex media by this treatment. The treated medium was used for growth studies in a gram-positive and a number of gram-negative organisms that were isolated from infections in humans. High-molecular-weight outer membrane proteins that are known to be induced under iron-depleted growth conditions (iron-regulated membrane proteins) were observed when a number of gram-negative pathogens were cultivated in the treated media. Iron uptake by Staphylococcus aureus varied, depending on the iron content of the medium.

Protein antigens of encapsulated Klebsiella pneumoniae surface exposed after growth in the presence of subinhibitory concentrations of cephalosporins.

It recently has been reported by us that cephalosporins, at a concentration below that influencing growth rate, reduced the production of enterochelin and capsule formation of iron-depleted Klebsiella pneumoniae. We now report on the antigenicity of the outer membrane components and surface-exposed protein antigens of iron-depleted cells grown in the presence or absence of cephalosporins. All major outer membrane proteins, including iron-regulated membrane proteins, were immunogenic. Encapsulated K. pneumoniae grown in antibiotic-free media had three protein antigens (60, 35.5, and 32.5 kilodaltons) exposed on the surface that were accessible to antibodies. Growth of the same cultures in the presence of subinhibitory concentrations of cephalosporins resulted in the exposure of a greater number of protein antigenic determinants, including iron-regulated membrane proteins, which become readily accessible to antibodies. It was also found that immunoblotting was generally more sensitive than conventional staining of the acrylamide gel with Coomassie blue in the detection of proteins.

In vivo evidence that bacteria in urinary tract infection grow under iron-restricted conditions.

The outer membrane protein composition of bacteria isolated directly and without subculturing from the urine of two patients with urinary tract infections was investigated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The results indicated that the bacteria grew under iron-restricted conditions, as revealed by the expression of several high-molecular-weight outer membrane proteins which could also be observed when the same isolates were grown under iron-depleted conditions in laboratory media. The antigenicity of outer membrane components of the bacteria isolated was studied by immunoblotting with serum samples from the patients. The results indicated that the sera from the patients contained antibodies against major outer membrane components of the bacteria present in the urine, including the iron-regulated membrane proteins.

Effect of subinhibitory concentrations of cephalosporins on surface properties and siderophore production in iron-depleted Klebsiella pneumoniae.

Subinhibitory MICs (sub-MICs) of several cephalosporins significantly reduced the enterochelin production of Klebsiella pneumoniae 327 grown under iron-depleted conditions and also reduced capsule formation regardless of iron availability. The surface hydrophobicity of K. pneumoniae 327 increased significantly when the bacteria were grown in either iron-sufficient or iron-depleted media in the presence of sub-MICs of all the cephalosporins used in this study. Antisera raised against a non-encapsulated K. pneumoniae strain caused rapid agglutination of K. pneumoniae 327 grown in the presence of sub-MICs of the cephalosporins but no agglutination of the same strain grown in drug-free media. The results indicated that the cephalosporins reduced enterochelin production and also capsule formation to the extent that noncapsular surface antigens were exposed, with possible significant consequences in vivo.