Gonococcal pilus vaccine. Studies of antigenicity and inhibition of attachment.

A gonococcal pilus vaccine or placebo was injected subcutaneously or intramuscularly into 71 human volunteers. The vaccine was found to be safe. The principal adverse reaction was a complaint of a sore arm, which was caused, at least in part, to the volume of material injected. 6 of 64 (9%) volunteers receiving the larger doses also complained of malaise. The vaccine was found to be antigenic. All of the volunteers developed an immunoglobulin class-specific antibody response as measured by a solid phase radioimmunoassay. The antibody was capable of blocking the attachment of gonococci to epithelial cells. A slight antibody response was also demonstrated to gonococcal lipopolysaccharide but the antibody responsible for blocking attachment of gonococci was directed entirely at the pilus protein. The stimulated antibodies were shown to crossreact with isolated pili of heterologous gonococcal strains and to block the attachment of heterologous gonococci. Absorption of immune sera by a heterologous pilus reduced the inhibition of attachment antibodies to pre-immune level, suggesting that the immune response was directed at a common pilus determinant.

Specific interaction of the tetragonally arrayed protein layer of Bacillus sphaericus with its peptidoglycan sacculus.

Tetragonal layer protein (T-layer) isolated from Bacillus sphaericus NTCC 9602 (wild type) or 9602 Lmw (variant) bonded specifically to the sacculi (peptidoglycan) of either cell type. Only uncleaved T-layer subunits were capable of specific recognition of the B. sphaericus sacculi; other Bacillus strains and gram-positive bacterial sacculi would not adsorb B. sphaericus strain 9602 T-layer. The peptidogylcan did not function as a template since isolated T-layer subunits self-assembled into characteristic pattern. Upon reassociation with sacculi, T-layer assemblies were randomly oriented patches compared with more continuous strictly oriented pattern on cells or fresh cell walls. T-layer associated with the sacculus was less susceptible to conditions that dissociated in vitro-assembled T-layer. Mild proteolysis of both wild-type and variant T-layer subunits by a variety of enzymes reduced the molecular weight by 18,000 in all cases, indicating that one region of the molecule was particularly susceptible to cleavage. Subunits from which the minor fragment had been cleaved upon aging retained the capacity to assemble in vitro, but would no longer adsorb to sacculi. Thus, the ability of T-layer to form networks was separate from its ability to bind cell walls, and the 18,000-dalton piece of the T-layer polypeptide was necessary for attachment to the cell wall.

Isolation, characterization, and in vitro assembly of the tetragonally arrayed layer of Bacillus sphaericus.

The tetragonally arranged cell wall layer (T-layer) of Bacillus sphaericus NTCC 9602 was isolated and characterized. Parallel studies were made on a spontaneous variant of the wild-type strain which had a T-layer subunit of altered molecular weight. A purification method for the T-layers was devised which involved separation of the cell walls from the cytoplasmic contents, urea dissociation of the T-layer from the cell walls, removal of soluble contaminants by differential centrifugation, and finally selective adsorption of uncleaved subunits to sacculi. The purified subunits retained the capacity to form an assembly in vitro with the same lattice parameters as that observed on whole cells or cell walls and could readsorb to the cell walls from which they had been extracted. Both the wild-type and the variant subunits behaved as single, homogeneous polypeptide chains. Carbohydrate assay and isoelectric point determinations revealed that both subunit types were acidic glycoproteins. Values obtained for thebuoyant density, isoelectric point, and extinction coefficient differed minimally; major differences were observed in the molecular weight and the characterisitc width of cylinders formed by in vitro-assembled T-layer of the wild-type and variant. Assembled T-layer was subject to alkaline or acid dissociation and in acid titration dissociated at its isoelectric point.

Identification and characterization of E. coli type-1 pilus tip adhesion protein.

The type-1 pilus of Escherichia coli is the prototype of this class of hair-like, multimeric adhesive organelles. This pilus mediates adherence to mannose-containing receptors on mucosal epithelia and other cells. The type-1 pilus, in one of several serological variants, is expressed by nearly all E. coli strains, and its promotion of colonization by pathogenic bacteria and the protective effects of purified pilus vaccines suggest that it is important as a bacterial virulence factor. Both the adhesive function and the serological variation of the type-1 pilus have been attributed to the thousand or so pilin protein monomers making up the pilus rods. This idea has been contradicted by our earlier observations on an E. coli strain expressing adhesion-defective pili. More recent genetic evidence also indicates that auxiliary pilus proteins are required for adhesive function. We report here the identification of three previously undetected integral minor proteins on the type-1 pilus, and show that one of them is the receptor-binding adhesin. This protein is antigenically conserved among strains with different pilin serotypes and is located at the pilus tip.

Purification of the Escherichia coli type 1 pilin and minor pilus proteins and partial characterization of the adhesin protein.

Type 1 pili of Escherichia coli contain three integral minor proteins with apparent molecular weights (Mr) of 28,000 (28K protein), 16,500, and 14,500 attached to rods composed of Mr-17,000 pilin subunits (Hanson and Brinton, Nature [London] 322:265-268). We describe here an improvement on our earlier method of pilus purification, which gives higher yields and higher purity. Also reported are methods allowing fractionation of intact type 1 pili into rods of pure pilin and free minor proteins, as well as fractionation of the 28K tip adhesion protein from the 16.5K and 14.5K proteins. We have determined the amino acid composition and amino-terminal sequence of the adhesion protein. This sequence shows limited homology with the amino-terminal sequences of several E. coli pilins, including type 1.

Cloning and expression in Escherichia coli of LKP pilus genes from a nontypeable Haemophilus influenzae strain.

Nontypeable Haemophilus influenzae HF0295, isolated by aspiration from the middle ear of a patient with otitis media, expresses long, thick, and hemagglutinating pili of a single serotype (LKP1) on its surface. An intact pilus vaccine consisting of the LKP1 serotype protected chinchillas against experimental otitis media (C. C. Brinton, Jr., M. J. Carter, D. B. Derber, S. Kar, J. A. Kramarik, A. C. C. To, S. C. M. To, and S. W. Wood, Pediatr. Infect. Dis. J. 8:554-561, 1989; R. B. Karasic, D. J. Beste, S. C. M. To, W. J. Doyle, S. W. Wood, M. J. Carter, A. C. C. To, K. Tanpowpong, C. D. Bluestone, and C. C. Brinton, Jr., Pediatr. Infect. Dis. J. 8:562-565, 1989). The genes encoding LKP1 pili were cloned from a genomic library of the clinical strain as a 12.5-kilobase insert on a plasmid vector and inserted into Escherichia coli K-12. Transposon mutagenesis and deletion constructs mapped the pilus-coding region within a 7-kilobase region of insert DNA. The recombinant bacteria were found by electron microscopy to express pili morphologically similar to LKP1 pili. Purified pilus rods from the recombinant and its parental strain were composed of a single detectable protein with an apparent molecular weight of 27,500. Antibodies raised against LKP1 pili purified from H. influenzae immunologically reacted with pili from the recombinant bacteria. Pili from both strains also adhered to human erythrocytes and buccal cells with the same specificity.

Mechanical removal of F pili, type I pili, and flagella from Hfr and RTF donor cells and the kinetics of their reappearance.

The effect of mechanical agitation (blending) on the removal of F pili, type I pili, and flagella from Hfr (high-frequency recombinant) and resistance transfer factor (RTF) fi(+)Escherichia coli cells was studied by electron microscopy. The reduction in number and length of appendages was measured as a function of blendor speed under standard conditions of temperature, medium, cell density, and blendor configuration. F pili and flagella were removed within the same narrow range of blendor speeds. Type I pili were removed within a higher and broader range of speeds. The speed which reduced the average length of type I pili to 50% was 3.5 times the speed which reduced the average length of F pili to 50%. None of the speeds employed inhibited cell growth, viability, or the ability to produce cell appendages. The kinetics of reappearance of F pili and type I pili after removal by blending were also different. F pili grew out very rapidly, reaching 50% of their full length in 30 sec and their full length in 4 to 5 min. The number of attached F pili per cell also increased rapidly, reaching a constant value in 4 to 5 min. After 5 min, F pilus lengths were distributed around a modal value of about 1.2 mum, and the numbers of F pili per cell were distributed according to a Poisson distribution, with an average of 1.0 per cell. These reappearance kinetics, length distributions, and number distributions are consistent with a model of F-pilus outgrowth in which new F pili appear at random locations on the cell surface at an average rate of about once every 4 min, grow to their characteristic length in about 4 min, and then separate from the cell. F pili which had separated could absorb to the cells, leading to the presence of two classes of F pili on cells: those in the process of natural out-growth and those attached by absorption. Type I pili increased in length much more slowly than did F pili, although the fraction of cells having visible type I pili increased very rapidly after blending because of the large number of type I pili per cell. The fraction of flagellated cells increased even more slowly, reaching only 30% of the unblended fraction in 30 min. The application of blending spectra and reappearance kinetics to the identification of cell functions with surface structures is discussed.

Inhibition of bacterial conjugation by ribonucleic acid and deoxyribonucleic acid male-specific bacteriophages.

Both deoxyribonucleic acid (DNA) and ribonucleic acid (RNA) male-specific phages, with an F-specific host range, inhibited the bacterial mating process of Escherichia coli. DNA phages prevented the formation of mating pairs but had no effect on mating pairs once they were formed. A step in RNA phage infection, prior to RNA penetration, prevented the formation of mating pairs and, in addition, prevented a fraction of existing mating pairs from completing the mating process. These findings are compatible with the hypothesis that donor cells have a single surface structure involved in both conjugation and male-phage adsorption and that this element is the F pilus.

Iminobiotin affinity columns and their application to retrieval of streptavidin.

A method is described for the retrieval of streptavidin from the culture broth of Streptomyces avidinii. The key step in this procedure is the adsorption of streptavidin from culture concentrates to an affinity column in which iminobiotin is attached to AH-Sepharose 4B. This column binds streptavbidin at pH 11 and releases the protein at pH 4. The recovery of streptavidin is practically quantitative. The pH dependence of the iminobiotin-avidin affinity, discovered by Green [Green, N. M. (1966) Biochem. J. 101, 774-779], has thus found practical application. The streptavidin bound 4.07 +/- 0.02 mol of [14C]biotin per mol and was essentially homogeneous as judged by disc and slab gel electrophoresis. Streptavidin was extensively succinoylated without loss of biotin-binding capacity. The observations that 125I-labeled streptavidin and 125I-labeled succinoylstreptavidin are retained by iminobiotin-AH-Sepharose 4B columns at pH 7.5 and are eluted at pH 4.0 provides a convenient purification method for these iodinated proteins. The technique employed for the retrieval of streptavidin is generally applicable to the isolation of iminobiotinylated molecules.

Genetic complementation analysis of Escherichia coli type 1 somatic pilus mutants.

A genetic complementation analysis of 75 stable nonpiliated mutants of a type 1 piliated strain of Escherichia coli K-12, AW405, was performed. Strains containing pairs of pil mutations were constructed by the infectious transfer of an F101 plasmid containing one pil mutation into E. coli K-12 AW 405 containing another pil mutation. The presence or absence of type 1 pili on the merodiploid strains was determined by agglutination with type 1 pilus antiserum. All 75 mutants fell into one of four complementation groups. The pattern of complementation defined three cistrons involved in pilus formation, pilA, pilB, and pilC. The fourth complementation group was composed of a large number of mutants defective in both pilA and pilB functions.

In vitro adhesion of Escherichia coli to porcine small intestinal epithelial cells: pili as adhesive factors.

Escherichia coli strains with pili (K99 or 987P) known to facilitate intestinal colonization adhered in vitro to porcine intestinal epithelial cells. These strains adhered equally to both ileal and jejunal epithelial cells. A laboratory E. coli strain that has type 1 pili also adhered to porcine intestinal epithelial cells. When nonpiliated cells derived from 987P+, K99+, or type 1 pilus+ strains were used for in vitro adhesion assays, they failed to adhere. The attachment of piliated bacteria to epithelial cells was a saturable process that plateaued at 30 to 40 bacterial cells attached per epithelial cell. Competitive inhibition of bacterial cell attachment to epithelial cells with purified pili showed that only purified 987P competed against the 987P+ strain and only purified type 1 pili competed against the type 1 pilus+ strain. Competition between a K99+ strain and K99 was not consistently achieved. K99+, 987P+, and type 1 pilus+ bacteria could be prevented from adhering to epithelial cells by Fab fragments specific for K99, 987P, or type 1 pili, respectively. Fab fragments specific for non-K99 bacterial surface antigens did not inhibit adhesion of the K99+ strain. It is concluded that adhesion of E. coli to porcine intestinal epithelial cells in vitro is mediated by pili and that the epithelial cells used apparently had different receptors for different pili.

Functions of F pili in mating-pair formation and male bacteriophage infection studies by blending spectra and reappearance kinetics.

The extent of removal at various blending speeds (blending spectrum) and the kinetics of reappearance after blending of the ability of male Escherichia coli bacteria to form mating pairs, to adsorb and be infected by ribonucleic acid male phage, and to adsorb and be infected by deoxyribonucleic male phage were identical to the blending spectrum and reappearance kinetics of microscopically visible F pili. The same results were obtained with an Hfr (high-frequency recombinant), F', or resistance transfer factor (RTF) fi(+) mating system. Blending did not affect the viability, growth rate, ability to adsorb T4 phage, or ability to produce new F pili at any of the speeds used. It can be concluded that microscopically visible F pili are an absolute requirement for all three functions. Three classes of F pili have been found in bacterial cultures: attached, adsorbed, and free. Bacteria with adsorbed F pili in addition to attached ones were proportionately more susceptible to male phage infection, suggesting that adsorbed F pili may be at least partially functional. Free F pili did not compete with bacteria for phage. Some implications of the virus-like nature of F-pilus outgrowth for the mechanisms of mating and male phage infection are discussed.

Immunization of suckling pigs against enteric enterotoxigenic Escherichia coli infection by vaccinating dams with purified pili.

Pregnant swine (gilts) were vaccinated parenterally with a suspension of purified pili from the porcine enterotoxigenic Escherichia coli strain 987 (09:K103::NM). Gilts injected with placebo served as controls. Suckling pigs born to gilts in both groups were challenged intragastrically with virulent strain 987. The percentage of deaths, incidence and duration of diarrhea, numbers of E. coli in the ilea, and E. coli attachment to the villous epithelia were significantly less in suckling pigs of vaccinated gilts than in those of controls. These results are consistent with the hypothesis that pili of some enterotoxigenic E. coli facilitate adhesion to intestinal epithelia. Vaccination of dams with pili appears to be a means of immunizing against diarrheal disease caused by enterotoxigenic E. coli in suckling neonates. This work confirms the role of somatic pili as colonization and virulence factors and provides another example of safe and effective purified pilus vaccines.

Antigenic specificity of antibodies in vaginal secretions during infection with Neisseria gonorrhoeae.

Antibodies in genital secretions of patients with gonorrhea have been shown to inhibit the attachment of gonococci to epithelial cells. The gonococcal antigens for which these antibodies are specific were studied by adsorption of the genital secretions from a patient infected with gonorrhea with purified lipopolysaccharide, outer membrane complex, or purified pili of homologous Neisseria gonorrhoeae and measurement of the reduction of inhibition of attachment of the gonococci to epithelial cells. The removal of antibodies was documented with the use of a solid-phase radioimmunoassay in which the amount of antibody in the adsorbed secretions that bound to a specific gonococcal antigen was shown to be reduced as compared with the amount of antibody in unadsorbed secretions. The antibody in the secretions that inhibited attachment was removed primarily by adsorption with the homologous pili, not with homologous lipopolysaccharide. A preparation of the homologous outer membrane complex that contained pili, cell-wall proteins, and lipopolysaccharide also blocked the inhibitory antibody.

Isolation and characterization of Escherichia coli phase variants and mutants deficient in type 1 pilus production.

Type 1 pili of Escherichia coli are the prototype of the somatic class of pili found on many strains of bacteria. As a first step in the genetic analysis of type 1 piliation, an extensive series of nonpiliated derivatives of E. coli K-12 strain AW405, was characterized to produce attached or free pili when examined in the antiserum or appeared to produce attached or free pili when examined in the electron microscope. The derivatives fell into two classes; phase variants and mutants. Phase variants that formed colonies of two distinctive types, one associated with a predominantly piliated (P+), and the other associated with a nonpiliated (P-) phase, were obtained. Each phase could give rise to the other at a relatively high rate, which was greater in the P- to P+ direction during culture in unshaken liquid medium. In addition, 77 Pil- mutants were selected on the basis of a subtle difference in colonial morphology. The mutants reverted, if at all, at a much lower rate than that of the P- to P+ change. The stability of Pil- derivatives grown in unshaken liquid medium was used as a criterion for distinguishing between phase variants and mutants, Phase variation also effected colonial morphology and chemotactic swarming. These properties did not directly depend upon piliation since Pil- mutants were only slightly altered in colonial form and unaltered in chemotactic swarming. Piliation of Pil+ bacteria was quantitatively affected by growth conditions.