Although Escherichia coli strains possessing the Kl capsule are predominant among isolates from neonatal E. coli meningitis and most of these Kl isolates are associated with a limited number of 0 lipopolysaccharide (LPS) types, the basis of this association of Kl and certain 0 antigens with neonatal E. coli meningitis is not clear. The present study examined in experimental E. coli bacteremia and meningitis in newborn and adult rats whether or not the Kl capsule and/or 0-LPS antigen are critical determinants in the development of meningitis. Rats received subcutaneously a Kl E. coli strain (018+K1+) or mutants lacking either the Kl capsule (018+K1-) or 0 side-chain (018-K1+). 12-24 h later, blood and cerebrospinal fluid (CSF) specimens were obtained for quantitative cultures. The isolation of E. coli from CSF was observed in both newborn and adult rats infected with K1+ strains regardless of LPS phenotype (018+ or 18-) who also developed a high degree of bacteremia (e.g., > 104 CFU/ml of blood). In contrast, none of the newborn and adult rats infected with 018+K1-and developing bacteremia of > 10' were found to have positive CSF cultures.These findings indicate that the presence of the Kl capsule and a high degree ofbacteremia are key determinants in the development of E. coli meningitis, suggesting that there may be specific binding sites present in the brain which have an affinity for the Kl capsule and thus may be responsible for the entry of Kl-encapsulated E. coli into the meninges. (J. Clin. Invest. 1992. 90:897-905.)

S U M M A R YNon-smooth mutants of Salmonella typhimurium strain L T~ with known lipopolysaccharide (LPS) defects were tested for sensitivity to smooth-specific phages (P22 and P22h and a newly isolated phage, 9NA, active on P22 lysogens); Felix 0 phage; and several rough-specific phages including C21. Smooth strains were sensitive only to the smooth-specific and Felix 0 phages. Six rfb mutants (unable to make 0 chains) were sensitive to Felix 0 and all the rough-specific phages except C21 (pattern R-sens). Of nine rfa mutants (presumed to have LPS core defects) four were R-sens, six were resistant to Felix 0 and some rough-specific phages (pattern R-res-I), and one was also resistant to phage Br2 (pattern R-res-a). The phage sensitivity of phosphomannoisomerase (prni) mutants was the same as that of rfb mutants, except that they were partly sensitive to P22h. UDPgalactose-epimerasenegative mutants were sensitive to C21 and various rough-specific phages including Br2 (pattern Epi-I). An rfc mutant (unable to polymerize 0 repeat units) was sensitive only to Felix 0 and P221 (pattern Zsr). A part-rough mutant of class D (with abnormally few 0 chains) was incompletely resistant to smooth-specific phages, resistant to Felix 0 but sensitive to all rough-specific phages except C21 (pattern D-I).Spontaneous and mutagen-induced non-smooth mutants were isolated from LT2 strains with appropriate markers by selection with Felix 0 and/or P22 phage. (One parent strain used was non-lysogenic for Fels 2, for which LT2 wild-type is lysogenic. Lysogeny for Fels 2 did not affect sensitivity to the other phages.) Some mutants gave new sensitivity patterns. Mutants of these and of previously unmapped classes were crossed with smooth Hfr strains. The rfc loci of two mutants and pmi loci of two others were located in the gal-trp segment. Three mutants of pattern R-sens yielded 0-specific hapten but mapped near his; they are believed to be unable to transfer 0 chains from antigen carrier lipid to the LPS core as a result of mutation at rfbT. Six mutants of pattern R-sens were smooth in cultural and serological properties; they mapped near his and are probably leaky rfb mutants. Many mutants had the class D part-rough phenotype, divisible by phage sensitivities into patterns D-I, D-2 and D-3. Mutants of all three classes mapped near xyl; they are likely to be rfa mutants, perhaps leaky, with LPS core defects which hinder but do not prevent attachment of 0 chains. Two classes were sensitive to C21 (Wilkinson & Stocker, 1968): rfaH mutants, of pattern Epi-I, unable to add the main-chain galactose unit of the core; and rfaG mutants, resistant to Br2 (pattern Epi-a), unable to add the proximal glucose unit. Both loci mapped in or near the strA-xyl-metA segment. Several non-smooth mutants did not grow in the presence of bile-salts. Three mutants (rfaF) made LPS deficient of the distal heptose unit; one mutant (rfaE) was unable to add the proximal heptose unit. Both these loci mapped in or near the strA-metA segment. I N T R O D U C T I O N...

Studies of the virulence and plasmid properties of Yersinia enterocolitica revealed a plasmid, 42.2 ± 1.1 megadaltons in size, which is associated with the pathogenicity and calcium dependency of this organism.

No abstract

Tumor necrosis factor/cachectin (TNF/C) is the principal mediator of bacterial endotoxin-induced shock and death. We found that the C3H/HeJ mouse, which is less able to produce TNF/C in response to endotoxin, has a 1,000-fold greater susceptibility to lethal infection with Escherichia coli than the TNF-responsive congenic mouse, C3H/HeN. This surprising finding suggested that this lethal peptide may also be involved in host protection. To test this hypothesis we pretreated the C3H/HeJ mouse with a combination of recombinant murine TNF/C-alpha and IL-1 alpha. This combination protected these mice against an intraperitoneal bacterial challenge of greater than 20 LD50S (nearly 2 x 10(2) CFU) that grew to a level of greater than 10(7) CFU/ml of blood and per gram of liver in untreated mice. This suggests a significant role for these cytokines in host defenses against invasive infections that require bacterial replication within the host. These protective mechanisms may not be important for less virulent organisms. These findings may have important implications for the proposed use of anti-TNF/C agents in the treatment of septic shock.

We examined 534 clinical isolates of Escherichia coli for sensitivity to rough lipopolysaccharide-specific and K1-specific phages. Twenty-eight percent of bacteremic isolates were sensitive to rough-specific phages. Forty-two percent of these strains, against only 20% of bacteremic isolates insensitive to rough-specific phages, had K1 capsule (P less than 0.001). K1-positive strains were usually resistant to phagocytic killing, whereas strains lacking the K1 capsule were more likely to be killed regardless of capsular type. Eighty-two percent of strains were typable with O-specific, 57% with K-specific, and 74% with H-specific antisera. Sixty percent of E coli were agglutinated by only 10 O-specific antisera. K1 was the most common capsular type, followed by K5, K2, and K12, whereas four H antigens accounted for nearly half of the H-typable strains. We conclude that (1) the combination of rough-specific and K1-specific phage sensitivity defines functionally similar groups of bacteria and (2) a polyvalent vaccine against invasive E coli is possible given the relatively limited number of invasive O:K:H serotypes.

Pili comprise several types of morphologically similar thin appendages growing out from the surface of gram-negative bacteria.'-3 Type I pili are composed of protein subunits of molecular weight 17,000 polymerized into rigid right-handed helices of diameter 70 A and pitch 24 A, having an axial hole 20-25 A in diameter.4 Other types of pili with different external diameters exist2 but their composition and fine structure is unknown.It had been considered previously that pili might be involved in the fertility of male bacteria.3 This speculation was based on the frequent occurrence of pili on male E. coli K 12 strains and on the plausibility of chromosomal transfer being mediated by a rod-like structure which could actively traverse the cell membrane and wall. However, no correlation of piliation with maleness was found. Many female strains were richly piliated and some cultures of male strains contained only a few sparsely piliated cells. The hypothesis could not be disproved, however, since none of the many male strains subsequently examined were ever completely nonpiliated. Although the majority of cells in a culture may have had no pili, at least a few cells could always be found with a few attached pili.A new method of approach to chis problem has been provided by the recent electron microscopic observations of Crawford and Gesteland5 who noted that a male-specific bacteriophage, R-17, adsorbed to pili of an Hfr and an F+ strain of E. coli but not to pili of an F-strain.We have investigated the adsorption of another male-specific bacteriophage, M 12, isolated by P. H. Hofschneider.8 Our studies revealed that M 12 adsorbs to some of the pili present on male bacteria and that phage adsorption can be used to distinguish them in electron micrographs from other types of pili occurring on the same cell. It is possible to demonstrate that these "F pili" are genetically controlled by the fertility factor of E. coli K 12.Materials and Methods.-Phage: M 12 phage, a small (about 270 A diameter), spherical, RNA-containing phage infecting Hfr and F+ strains but not F-strains, was obtained from Dr. P. H. Hofschneider.Electron microscopy: Bacteria were grown in tryptone yeast extract calcium broth (10 gm tryptone, 5 gin yeast extract, 5 gm NaCl, 0.75 gm CaCl2 2H20 per liter of water, the CaC12 sterilized separately). Overnight, unshaken, unaerated cultures were diluted 1:10 into fresh medium and grown for 3-4 hr under the same conditions until the bacterial density was about 5.108 cells per ml. M 12 phage was added at multiplicities from 5 to 100 and the mixture incubated at 370C. After 10 min, the mixture was rapidly chilled in an ice bath and prepared for the electron microscope by the collodion agar filtration method of Kellenberger.7 Tests for phage susceptibility: (A) Plaque formation: Standard phage plaque methods using tryptone yeast extract calcium soft agar (0.7%) and bottom agar 776

Extraintestinally invasive Escherichia coli (EC) that possess both a complete LPS and K1 capsule evade both complement-mediated bacteriolysis and neutrophil-mediated killing. Since C3H/HeJ mice that are hyporesponsive to LPS were uniquely susceptible to lethal infection with EC of this phenotype, we speculated there was an LPS-initiated host defense mechanism against this pathogenic phenotype. The LPS-normoresponsive C3H/HeN as well as the C3H/HeJ mice cleared these EC from the circulation within 4 h of intravenous administration. Whereas electron micrographs of the liver demonstrated these EC undergoing degeneration within the phagolysosomes of of both macrophages and Kupffer cells of C3H/HeN mice, these EC replicated within these cells of the C3H/HeJ mice. Restoration of anti-EC activity of C3H/HeJ mice occurred with activation of Kupffer cells and peritoneal macrophages in vivo with BCG and in vitro with IFN-y, but not with LPS. Pretreatment of C3H/ HeJ mice with a combination of recombinant murine IL-1 and TNF-a also restored the killing of K1 +-EC but did not enhance the killing of a K1--EC mutant. These data are consistent with the hypothesis that (a) there is no intrinsic inability of C3H/HeJ phagocytes to kill EC, but (b) an LPS-initiated, cytokine-mediated host defense mechanism is required for such killing. These studies emphasize the importance of bacterial surface characteristics in the interaction with specific host defenses. (J. Clin. Invest. 1995.