Klebsiella pneumoniae in orange juice concentrate

Fecal coliform-positive, capsule-forming Klebsiella pneumoniae cells were observed in high densities (10(4) to 10(8) CFU/100 ml) in two commercial batches of frozen orange juice concentrate at a cannery in Puerto Rico. Contamination of both lots was gross and included off colors and odors. Isolates of K. pneumoniae from these concentrates revealed growth at 4, 25, and 34 degrees C with generation times from 0.39 to 1.84 h.     

Characterization of the koala biovar of Chlamydia pneumoniae at four gene loci--ompAVD4, ompB, 16S rRNA, groESL spacer region

Koalas are infected with two species of Chlamydia, C. pecorum and C. pneumoniae. While it is known that significant genetic diversity occurs in the C. pecorum strains infecting koalas, very little is known about the C. pneumoniae strains that infect this host. In the current study, 10 isolates of koala C. pneumoniae were analysed at four gene loci and found to be different to both the human and horse C. pneumoniae strains at all loci (biovar differences ranging from 0.3% at groESL up to 9.0% at ompAVD4). All koala biovar isolates studied were found to be 100% identical at ompAVD4 (all 10 isolates) and at ompB (all three isolates) gene. This lack of allelic polymorphisms at ompAVD4 has now been observed for koala C. pneumoniae, human C. pneumoniae, guinea pig inclusion conjuctivitis C. psittaci and feline conjuctivitis C. psittaci and may be correlated to a lack of antibody response to the chlamydial major outer membrane protein (MOMP) in these same strain/host combinations. This study also provides the first documented case of natural C. pneumoniae infection causing a severe and extended respiratory episode in a captive koala population. This captive episode is in contrast to most free-range observations in which koala C. pneumoniae is rarely documented as causing respiratory, ocular or urogenital tract disease.     

High-yield culture and purification of Chlamydiaceae bacteria

Research on intracellular bacteria of the family Chlamydiaceae, and the diseases they cause, requires large amounts of infectious elementary bodies (EB). We describe an approach that maximizes the generation of Chlamydia pneumoniae, Chlamydia trachomatis, Chlamydia abortus, or Chlamydia pecorum EBs in several replication cycles over approximately 10 days or more in a saturated equilibrium monolayer cell culture system. Buffalo Green Monkey Kidney (BGMK) cells, Human Epidermoid Carcinoma-2 (HEp-2) cells, or mouse McCoy cells were tested. BGMK cells best supported C. pneumoniae replication when cultivated in Iscove's Modified Dulbecco's Medium. From day 1 to day 9 after inoculation, C. pneumoniae genomes per ml culture medium increased from 10(5.1) to 10(8.6) in BGMK, from 10(5.6) to 10(8.1) in HEp-2, and remained at 10(5.2) in McCoy cell cultures. Three-month pre-inoculation maintenance of BGMK cells in different culture media did not influence C. pneumoniae yields. Inoculation at multiplicities of infection (MOI) of 10 or higher and supplementation of the cell culture medium on day 7 after inoculation with 0.1% glucose enhanced C. pneumoniae EB yields in harvested cell culture medium. For purification, EBs in medium were concentrated by sedimentation, followed by low-speed centrifugation for removal of host cell nuclei, and by step-gradient centrifugation of the supernatant in a 30% RenoCal-76-50% sucrose step-gradient. Extensive sonication increased yield and infectivity of chlamydial EB. The combined method typically produced from 1000 ml infected BGMK culture medium 10 ml homogeneous, single-cell, highly infectious EB stock containing approximately 5x10(11) C. pneumoniae genomes equivalent to 4-5x10(11) inclusion forming units.     

Infection of Acanthamoeba castellanii by Chlamydia pneumoniae.

Chlamydia pneumoniae is an intracellular respiratory pathogen, which, similar to Legionella, might have developed mechanisms to escape the intracellular bactericidal activity of both human host cells and amoeba. We therefore investigated the intracellular growth and survival of C. pneumoniae in Acanthamoeba castellanii by using cell culture, immunofluorescence microscopy, and electron microscopy. A castellanii was incubated with purified elementary bodies of C. pneumoniae TW 183 at a concentration of 10(6) inclusion-forming units (IFU)/ml to give a ratio of approximately 1 IFU of C. pneumoniae per amoeba. Quantitative determination of chlamydial growth within A. castellanii revealed viable and infective C. pneumoniae in the range of 10(4) to 10(5) IFU/ml between days 7 and 14 postinfection. Immunofluorescence analysis and transmission electron microscopy with subsequent immunogold staining confirmed evidence of infection of the amoebae by C. Pneumoniae and additionally revealed that C. pneumoniae entered the typical growth cycle. Our results show that amoebae allow the survival of C. pneumoniae, suggesting that amoebae may serve as an additional reservoir for Chlamydia or Chlamydia-related organisms.     

Reduction of cyclopropene by NifV- and wild-type nitrogenases from Klebsiella pneumoniae.

The nitrogenase from wild-type Klebsiella pneumoniae reduces cyclopropene to cyclopropane and propene in the ratio 1:2 at pH 7.5. We show in this paper that the nitrogenase from a nifV mutant of K. pneumoniae also reduces cyclopropene to cyclopropane and propene, but the ratio of products is now 1:1.4. However, both nitrogenases exhibit the same Km for cyclopropene (2.1 x 10(4) +/- 0.2 x 10(4) Pa), considerably more than the Km for the analogous reaction with Azotobacter vinelandii nitrogenase under the same conditions (5.1 x 10(3) Pa). Analysis of the data shows that the different product ratio arises from the slower production of propene compared with cyclopropane by the mutant nitrogenase. During turnover, both nitrogenases use a large proportion of the electron flux for H2 production. CO inhibits the reduction of cyclopropene by both K. pneumoniae proteins, but the mutant nitrogenase exhibits 50% inhibition at approx. 10 Pa, whereas the corresponding value for the wild-type nitrogenase is approx. 110 Pa. However, H2 evolution by the mutant enzyme is much less affected than is cyclopropene reduction. CO inhibition of cyclopropene reduction by the nitrogenases coincides with a relative increase in H2 evolution, so that in the wild-type (but not the mutant) the electron flux is approximately maintained. The cyclopropane/propene production ratios are little affected by the presence of CO within the pressure ranges studied at least up to 50% inhibition.     

Etiology of acute bronchitis in young children

In the microbiological examination of 132 children aged 0-3 years with acute bronchitis, Streptococcus pneumoniae and Haemophilus influenzae at a concentration of greater than or equal to 10(4) cells/ml, as well as different species of opportunistic bacteria, were isolated from tracheobronchial washings obtained from 100 of these children. S. pneumoniae and H. influenzae were found to play the leading role in the etiology of the acute bacterial inflammatory process in acute bronchitis in children.     

Use of PCR methods for detection of selected genes of Mycoplasma pneumoniae

The aim of this study was standardization of PCR for the detection of gene encoding the P1 protein, 16S rRNA and elongation factor Tu of M. pneumoniae. A total of 13 strains of M. pneumoniae, 28 strains of other mycoplasmas and 14 strains of different bacteria causing respiratory tract infections were tested. In all of tested M. pneumoniae strains the presence of the sought genes was confirmed. The specificity of DNA was confirmed by the restriction endonuclease analysis with enzymes Hind III, Alu I and Hha I. With none of primers specific for the M. pneumoniae genes amplification of DNA from other bacteria was noted. The PCR method with the selected primers allowed to detect from 10(2) to 10(4) cfu M. pneumoniae/ml suspended in broth. The obtained results indicate that the PCR method can be used for detection of M. pneumoniae genes. A very good sensitivity and specificity predestine++ PCR as a potential quick diagnostic method for identification of M. pneumoniae in clinical specimens.     

Short and long courses of ofloxacin therapy of Klebsiella pneumoniae sepsis following irradiation.

Exposure to whole-body irradiation is associated with fatal gram-negative sepsis. The optimal length of therapy of such infection is not established. The effect of short and long courses of oral therapy with the quinolone ofloxacin for orally acquired Klebsiella pneumoniae infection was tested in B6D2F1 mice exposed to 8.0 Gy of bilateral radiation from 60Co. A dose of 10(8) organisms was given orally 4 days after irradiation, and therapy was started 1 day later. Cultures of the ileum 7 days after irradiation showed the recovery of K. pneumoniae in 7 of 10 untreated mice and in 3 of 20 treated with ofloxacin. However, 14 days after irradiation K. pneumoniae was isolated in 5 of 6 untreated mice, in 7 of 9 that received the short course of therapy, and in one of those that received the long course of therapy (P less than 0.05). At Day 7, K. pneumoniae was isolated from the livers of 6 of 10 untreated mice, and from none of those receiving ofloxacin (P less than 0.05). At 14 days, K. pneumoniae was isolated in 4 of 6 untreated animals, in 4 of 9 that received the short course of therapy, and in none of the mice that received the long course of therapy (P less than 0.05). Only 3 of 20 (15%) untreated mice survived for 30 days as compared to 11 of 20 (55%) mice treated for 7 days with ofloxacin and 18 of 20 (90%) mice treated for 21 days with ofloxacin (P less than 0.05). These survival data illustrate the efficacy of a 21-day course over a 7-day course of ofloxacin therapy for orally acquired K. pneumoniae infection in irradiated hosts.     

Nitrogen-deficient Medium in the Differential Isolation of Klebsiella and Enterobacter from Feces

On a nitrogen-deficient agar medium, the tribe Klebsielleae formed large, glistening, mucoid colonies which were easily distinguished from other colony types. Of 113 Klebsielleae isolates from human feces which were characterized, Klebsiella accounted for 88% of the total; 75% were K. pneumoniae; K. ozaenae (13%) was isolated from one individual only. The remaining strains (12%) were identified as Enterobacter cloacae. Counts (for the tribe) ranged from 10(2) to 10(6), with a median of 10(4); 9 of 53 stool specimens were negative. K. pneumoniae was also isolated from 6 of 41 frozen foil-pack foods. Anaerobic studies at room temperature and 37 C revealed no appreciable differences from aerobic plates. The nitrogen-deficient medium appeared better than E M B for isolation of Klebsielleae when they were present in low numbers relative to other coliforms; slime production by Klebsielleae concomitant with minimal growth of other bacteria is involved.     

Source and extent of Klebsiella pneumoniae in the paper industry.

Three pulp and paper mill processing plants were evaluated for fecal coliform and Klebsiella pneumoniae bacterial concentrations. Freshwater consumed by paper industries contained minimum detectable levels of K. pneumoniae, less than 10 organisms per 100 ml. Elevated concentrations of K. pneumoniae could be traced from early pulping stages to water processing reuse systems. Concentrations of K. pneumoniae (thermotolerant and thermointolerant) ranged from 40,000 organisms per 100 ml to an estimated 3 x 10(6) organisms per 100 ml. K. pneumoniae biotyping provided evidence for the selective growth and persistence of K. pneumoniae from the initial wood washing stages through to the final effluent discharge. Wastewater treatment had limited effects in reducing K. pneumoniae concentrations. K. pneumoniae levels ranged from 40 organisms per 100 ml to an estimated 10(6) organisms per 100 ml. The presence of K. pneumoniae in water indicates degraded water quality, and its significance with regard to human health effects has yet to be examined.     

Adhesion of Mycoplasma pneumoniae to sulfated glycolipids and inhibition by dextran sulfate

A virulent strain of Mycoplasma pneumoniae was metabolically labeled with [3H]palmitate and studied for binding to glycolipids and to WiDr human colon adenocarcinoma cells. The organism binds strongly to sulfatide and other sulfated glycolipids, such as seminolipid and lactosylsulfatide which all contain terminal Gal(3SO4) beta 1-residues and weakly to some neolactoseries neutral glycolipids. M. pneumoniae do not bind gangliosides including the sialylneolacto-series and other neutral glycolipids that were tested. Only metabolically active M. pneumoniae cells bind to sulfatide, as binding is maximal in RPMI medium at 37 degrees C and almost completely abolished in nutrient-deficient medium or by keeping the cells at 4 degrees C. Dextran sulfate but not other sulfated or anionic polysaccharides at 10 micrograms/ml completely inhibits binding of M. pneumoniae to purified sulfatide. Dextran sulfate does not inhibit binding to the neolacto-series neutral glycolipids. Dextran sulfate partially inhibits adhesion of M. pneumoniae to cultured human colon adenocarcinoma cells (WiDr). The biological relevance of these data is suggested by our finding that sulfatide occurs in large amounts in human trachea, lung, and WiDr cells. Thus, there are at least two distinct receptors that mediate binding of M. pneumoniae to cells: glycolipids containing terminal Gal(3SO4) beta 1-residues as reported here, and glycoproteins containing terminal NeuAc alpha 2-3Gal beta 1-4GlcNAc sequences (Roberts, D. D., Olson, L. D., Barile, M. F., Ginsburg, V., and Krivan, H. C. (1989) J. Biol. Chem. 264, 9289-9293).     

A latex agglutination test for the detection of Mycoplasma pneumoniae in respiratory exudates: a comparative study with a commercially available DNA-probe test.

We prepared polyclonal antibody specific to Mycoplasma pneumoniae. Using this antibody, we developed a latex agglutination test (LAT) for detecting the organism in respiratory exudates as rapid diagnosis of M. pneumoniae infection. Further, LAT was compared with DNA-probe test (DP) which was the only commercially available test for the rapid detection of the organism. In LAT, both M. pneumoniae and M. genitalium give positive agglutination, but the titer of M. genitalium was significantly lower than that of M. pneumoniae. The detection limit of LAT was 2 x 10(5) CFU/ml and that of DP was 5 x 10(4) CFU/ml in vitro. It was considered that target molecules in LAT were accumulated in the pharyngeal portion of the patients, because of their long half-life at 37 C. However, ribosomal RNA which was target molecule in DP was destroyed at 37 C much sooner, and the accumulation could not be expected. Actually, positive rate in LAT was higher than that in DP among clinical specimens in which M. pneumoniae was detected by culture method. The procedure of LAT is much easier and more rapid than that of DP in which radioactive isotope is required. LAT could be the choice of test for rapid diagnosis of M. pneumoniae infection.     

IL-10 is an important mediator of the enhanced susceptibility to pneumococcal pneumonia after influenza infection

Secondary pneumococcal pneumonia is a serious complication during and shortly after influenza infection. We established a mouse model to study postinfluenza pneumococcal pneumonia and evaluated the role of IL-10 in host defense against Streptococcus pneumoniae after recovery from influenza infection. C57BL/6 mice were intranasally inoculated with 10 median tissue culture infective doses of influenza A (A/PR/8/34) or PBS (control) on day 0. By day 14 mice had regained their normal body weight and had cleared influenza virus from the lungs, as determined by real-time quantitative PCR. On day 14 after viral infection, mice received 10(4) CFU of S. pneumoniae (serotype 3) intranasally. Mice recovered from influenza infection were highly susceptible to subsequent pneumococcal pneumonia, as reflected by a 100% lethality on day 3 after bacterial infection, whereas control mice showed 17% lethality on day 3 and 83% lethality on day 6 after pneumococcal infection. Furthermore, 1000-fold higher bacterial counts at 48 h after infection with S. pneumoniae and, particularly, 50-fold higher pulmonary levels of IL-10 were observed in influenza-recovered mice than in control mice. Treatment with an anti-IL-10 mAb 1 h before bacterial inoculation resulted in reduced bacterial outgrowth and markedly reduced lethality during secondary bacterial pneumonia compared with those in IgG1 control mice. In conclusion, mild self-limiting influenza A infection renders normal immunocompetent mice highly susceptible to pneumococcal pneumonia. This increased susceptibility to secondary bacterial pneumonia is at least in part caused by excessive IL-10 production and reduced neutrophil function in the lungs.     

Streptococcus pneumoniae adhesion to mouse tracheal explants and its inhibition by carbohydrate preparations]

The method for the cultivation of tracheal explants of non-inbred mice was developed. The adhesive capacity of 34 S. pneumoniae strains belonging to 15 serotypes was studied. The geometric mean of S. pneumoniae concentrations attached to the organ culture was 8.93 x 10(4) CFU/ml. D-galactose blocked adhesion, on the average, 4.35 x 10(4) (95.5%) CFU/ml. The level of S. pneumoniae adhesion and the degree of its inhibition were dose- and strain-dependent. The use of different variants of the combination of D-galactose and S. pneumoniae made it possible to come to the conclusion that this monosaccharide blocked the receptors of epithelial cells and bacterial adhesins.     

Influence of temperature and relative humidity on the survival of Chlamydia pneumoniae in aerosols.

The survival of Chlamydia pneumoniae in aerosols was investigated by using a chamber with a capacity of 114.5 liters. We injected 5 x 10(7) inclusion-forming units (IFU) of C. pneumoniae in aerosols with a droplet size of 3 to 5 microns. Samples were taken after 30 s and every 1 min thereafter. The survival of C. pneumoniae was measured at four temperatures (8.5, 15, 25, and 35 degrees C) and at three different relative humidities (RH) of 5, 50, and 95% for each temperature. The survival rates of Streptococcus pneumoniae, Streptococcus faecalis, Klebsiella pneumoniae, Chlamydia trachomatis LGV2, and cytomegalovirus were also determined at 25 degrees C and 95% RH and compared with that of C. pneumoniae. At the mentioned temperatures and RH, a rapid decrease of C. pneumoniae IFU was observed in the first 30 s. After this the decrease in the number of IFU was more gradual. The survival of C. pneumoniae in aerosols were optimal at 15 to 25 degrees C and 95% RH; it was good compared with those of other microorganisms. A lower death rate was observed only in S. faecalis. In C. trachomatis, the death rate during the first 30 s was higher than that in C. pneumoniae (85 and 53.3%, respectively). After the first 30 s, the death rates in the two organisms were identical. It was concluded that transmission of C. pneumoniae via aerosols was possible. There is probably a direct transmission from person to person, taking into account the relatively short survival period of C. pneumoniae in aerosols.     

Construction of otherwise isogenic serotype 6B, 7F, 14, and 19F capsular variants of Streptococcus pneumoniae strain TIGR4

The polysaccharide capsule is the primary virulence factor in Streptococcus pneumoniae. There are at least 90 serotypes of S. pneumoniae, identified based on the immunogenicity of different capsular sugars. The aim of this study was to construct pneumococcal strains that are isogenic except for capsular type. Serotype 4 strain TIGR4 was rendered unencapsulated by recombinational replacement of the capsular polysaccharide synthesis (cps) locus with the bicistronic Janus cassette (C. K. Sung, J. P. Claverys, and D. A. Morrison, Appl. Environ. Microbiol. 67:5190-5196, 2001). In subsequent transformation with chromosomal DNA, the cassette was replaced by the cps locus derived from a strain of a different serotype, either 6B, 7F, 14, or 19F. To minimize the risk of uncontrolled recombinational replacements in loci other than cps, the TIGRcps::Janus strain was "backcross" transformed three times with chromosomal DNA of subsequently constructed capsular type transformants. Capsular serotypes were confirmed in all new capsule variants by the Quellung reaction. Restriction fragment length polymorphism (RFLP) analysis of the cps locus confirmed the integrity of the cps region transformed into the TIGR strain, and RFLP of the flanking regions confirmed their identities with the corresponding regions of the recipient. Transformants had in vitro growth rates greater than or equal to that of TIGR4. All four strains were able to colonize C57BL/6 mice (female, 6 weeks old) for at least 7 days when mice were intranasally inoculated with 6 x 10(6) to 8 x 10(6) CFU. The constructed capsular variants of TIGR4 are suitable for use in studies on the role of S. pneumoniae capsular polysaccharide in immunity, colonization, and pathogenesis.     

Transformation of a type 4 encapsulated strain of Streptococcus pneumoniae

Streptococcus pneumoniae strain JNR.7/87 is a highly virulent, type 4 encapsulated Gram-positive bacterium whose transformability has not been tested previously, and whose genome is currently being sequenced. The strain was transformed at very low efficiency by addition of exogenous competence-stimulating peptide: However, the efficiency was too low and irreproducible to be useful in many genetic studies. Therefore, the effects on transformation efficiency of changing different components of competence-stimulating peptide-induced transformation have been examined. Screening of growth media was followed by optimization of pre-induction culture acidification, glycine concentration, and induction time. An optimized protocol was developed whereby S. pneumoniae strain JNR.7/87 was transformed reproducibly with a streptomycin resistance (SmR) marker at an efficiency of approximately 10(5) colony forming units per 10(8) cells.     

Protective efficacy of a Mycoplasma pneumoniae P1C DNA vaccine fused with the B subunit of Escherichia coli heat-labile enterotoxin

In the present study, we investigated the immunomodulatory responses of a DNA vaccine constructed by fusing Mycoplasma pneumoniae P1 protein carboxy terminal region (P1C) with the Escherichia coli heat-labile toxin B subunit (LTB). BALB/c mice were immunized by intranasal inoculation with control DNAs, the P1C DNA vaccine or the LTB-P1C fusion DNA vaccine. Levels of the anti-M. pneumoniae antibodies and levels of interferon-γ and IL-4 in mice were increased significantly upon inoculation of the LTB-P1C fusion DNA vaccine when compared with the inoculation with P1C DNA vaccine. The LTB-P1C fusion DNA vaccine efficiently enhanced the M. pneumoniae-specific IgA and IgG levels. The IgG2a/IgG1 ratio was significantly higher in bronchoalveolar lavages fluid and sera from mice fusion with LTB and P1C than mice receiving P1C alone. When the mice were challenged intranasally with 10(7) CFU M. pneumoniae strain (M129), the LTB-P1C fusion DNA vaccine conferred significantly better protection than P1C DNA vaccine (P < 0.05), as suggested by the results, such as less inflammation, lower histopathological score values, lower detectable number of M. pneumoniae strain, and lower mortality of challenging from 5 × 10(8) CFU M. pneumoniae. These results indicated that the LTB-P1C fusion DNA vaccine efficiently improved protective efficacy against M. pneumoniae infection and effectively attenuated development of M. pneumoniae in mice.     

Susceptibility of irradiated B6D2F1/J mice to Klebsiella pneumoniae administered intratracheally: a pulmonary infection model in an immunocompromised host

Bacteria such as Klebsiella pneumoniae can invade and colonize an immunocompromised host and complicate clinical recovery. In the study reported here, an experimental model of induced pneumonia was developed in 60Co gamma-photon-irradiated mice for the purpose of evaluating efficacy of therapeutic agents. The model was characterized by use of probit analysis of bacterial dose, and microbiologic, and histopathologic results. Bacterial colony-forming-unit (CFU) values producing 50% mortality within 30 days (LD50/30) and their 95% confidence intervals were 4.0 x 10(4) [1.7 x 10(4) - 8.9 x 10(4)] for 0-Gray (Gy)-irradiated mice, 1.9 x 10(4) [7.0 x 10(3) - 4.8 x 10(4)] for 5-Gy-irradiated mice, and 1.0 x 10(3) [2.8 x 10(2) - 3.3 x 10(3)] for 7-Gy-irradiated mice. Probit regression line fits calculated by use of an iterative, weighted least-squares fit, were used to assess a dose-modifying factor (DMF). The DMFs for mortality, compared with that for the 0-Gy dose, with their 95% confidence intervals, were 2.2 [0.63 - 7.7] for the 5-Gy and 38.9 [9.6 -165.0] for 7-Gy doses. The 5-Gy probit line did not significantly differ (P = 0.21) from the 0-Gy probit line (dose ratios did not significantly differ from 1), whereas the 7-Gy probit line differed significantly from the 0-Gy probit line (P < 0.001). These results demonstrate that 7-Gy 60Co gamma-photon radiation in combination with intratracheal K. pneumoniae challenge induces a valid pulmonary infection model in immunocompromised female B6D2F1/J mice.     

Alternative prey: a mechanism for elimination of bacterial species by protozoa

Antibiotic-resistant strains of Salmonella typhimurium and Klebsiella pneumoniae died readily after their addition to raw sewage, but they grew in sterilized sewage. The decline was not a result of abiotic stresses, and because the bacteria were able to survive in large numbers for at least 15 days in solutions containing no organic nutrients, it was not a result of competition. Toxin production, bacteriophages, and Bdellovibrio sp. did not cause the disappearance of the two bacterial species. A decline was also evident if the sewage was first passed through a 3-micron (pore size) filter or treated with cycloheximide or cycloheximide plus nystatin, but protozoa developed under these conditions. Little or no decline occurred if the sewage was filtered and treated with the eucaryotic inhibitors before the addition of S. typhimurium or K. pneumoniae, and protozoa were not detected. S. typhimurium increased in abundance if cycloheximide, streptomycin, and erythromycin or large amounts of glucose were added to sewage. Tetrahymena thermophilus did not significantly reduce the population of S. typhimurium in buffer when the density of the bacterium was about 10(4)/ml. However, when more than 10(8) Enterobacter agglomerans cells per ml were added to the buffer, T. thermophilus reduced the abundance of E. agglomerans and S. typhimurium to 10(6) and 10/ml, respectively. The density of S. typhimurium was further decreased by a second increment of E. agglomerans cells. The disappearance of S. typhimurium and K. pneumoniae from sewage thus is the result of predation by protozoa.(ABSTRACT TRUNCATED AT 250 WORDS)