Survival and virulence of copper- and chlorine-stressed Yersinia enterocolitica in experimentally infected mice.

The effect of gastric pH on the viability and virulence of Yersinia enterocolitica O:8 after exposure to sublethal concentrations of copper and chlorine was determined in mice. Viability and injury were assessed with a nonselective TLY agar (tryptic soy broth containing lactose, yeast extract, and agar) and two selective media, TLYD agar (TLY agar plus sodium deoxycholate) and CIN agar (cefsulodin-Irgasan-novobiocin agar). Both copper and chlorine caused injury which was manifested by the inability of the cells to grow on selective media. CIN agar was more restrictive to the growth of injured cells than TLYD agar. Injury of the exposed cells was further enhanced in the gastric environment of mice. Besides injury, the low gastric pH caused extensive loss of viability in copper-exposed cells. Lethality in the chlorine-exposed cells was less extensive, and a portion of the inoculum (5.2 X 10(5) of 1 X 10(7) inoculated cells) reached the small intestine 5 min postinoculation. No adverse effect on the injured cells was apparent in the small intestine, and a substantial revival (approximately 70%) of the injury occurred in 3 to 4 h after intraluminal inoculation. The virulence of chlorine-stressed Y. enterocolitica in orally inoculated mice was similar to that of the control culture, but copper-stressed cells showed reduced virulence. Virulence was partly restored by oral administration of sodium bicarbonate before the inoculation of copper-exposed cells. Neutralization of gastric acidity had no effect on the virulence of the control or chlorine-stressed cells. The results of this study indicate that the extensive injury caused by the low gastric pH does not affect the virulence potential of chlorine-exposed cells. However, extensive cell death in the mouse stomach is responsible for the reduced virulence of the copper-stressed bacteria.     

Changes in virulence of waterborne enteropathogens with chlorine injury

We designed experiments to assess the effect of chlorine injury on the virulence of waterborne enteropathogens. Higher chlorine doses (0.9 to 1.5 mg/liter) were necessary to produce injured Yersinia enterocolitica, Salmonella typhimurium, and Shigella spp. than to produce injured enterotoxigenic Escherichia coli or coliform bacteria (0.25 to 0.5 mg/liter) in the test system used; 50% lethal dose experiments in which mice were used showed that injured Y. enterocolitica cells were 20 times less virulent than uninjured control cells (3,300 and 160 CFU, respectively). This decrease in virulence was not related to reduced attachment to Henle 407 intestinal epithelial cells, but could be related to a loss of HeLa cell invasiveness. In contrast, injured S. typhimurium and enterotoxigenic E. coli cells lost their ability to attach to Henle cells. These data show that some enteropathogens and coliform bacteria differ in their sensitivities to chlorine injury and that the virulence determinants affected by chlorine may vary from one pathogen to another.     

Changes in virulence of waterborne enteropathogens with chlorine injury.

We designed experiments to assess the effect of chlorine injury on the virulence of waterborne enteropathogens. Higher chlorine doses (0.9 to 1.5 mg/liter) were necessary to produce injured Yersinia enterocolitica, Salmonella typhimurium, and Shigella spp. than to produce injured enterotoxigenic Escherichia coli or coliform bacteria (0.25 to 0.5 mg/liter) in the test system used; 50% lethal dose experiments in which mice were used showed that injured Y. enterocolitica cells were 20 times less virulent than uninjured control cells (3,300 and 160 CFU, respectively). This decrease in virulence was not related to reduced attachment to Henle 407 intestinal epithelial cells, but could be related to a loss of HeLa cell invasiveness. In contrast, injured S. typhimurium and enterotoxigenic E. coli cells lost their ability to attach to Henle cells. These data show that some enteropathogens and coliform bacteria differ in their sensitivities to chlorine injury and that the virulence determinants affected by chlorine may vary from one pathogen to another.     

Reduced virulence of Yersinia enterocolitica by copper-induced injury.

A sublethal concentration of copper (0.75 mg/liter) caused substantial injury (87 to 95%) of Yersinia enterocolitica serotype O:8 cells in 72 h at 4 degrees C without producing extensive cell death. Copper-injured cells had a higher 50% lethal dose in mice (2,700 CFU) than uninjured cells (150 CFU). This reduced virulence correlated with more rapid clearance of the injured cells from the blood of mice after intravenous inoculation. A possible role of the liver in this process was shown by significant cell accumulation in mouse livers when copper-injured Y. enterocolitica cells were administered, compared with uninjured bacteria. In vitro studies with isolated mouse liver membranes showed higher titers of aggregation with copper-injured cells than control cells. The in vitro aggregation reaction and blood clearance activity in vivo were abolished by sugars that are known to interact with a hepatic lectin. Our data suggest that copper-induced injury reduces the virulence of Y. enterocolitica and that the liver may be involved in nonimmune rapid clearance of the injured cells, probably by interaction with a hepatic lectin(s).     

Reduced virulence of Yersinia enterocolitica by copper-induced injury

A sublethal concentration of copper (0.75 mg/liter) caused substantial injury (87 to 95%) of Yersinia enterocolitica serotype O:8 cells in 72 h at 4 degrees C without producing extensive cell death. Copper-injured cells had a higher 50% lethal dose in mice (2,700 CFU) than uninjured cells (150 CFU). This reduced virulence correlated with more rapid clearance of the injured cells from the blood of mice after intravenous inoculation. A possible role of the liver in this process was shown by significant cell accumulation in mouse livers when copper-injured Y. enterocolitica cells were administered, compared with uninjured bacteria. In vitro studies with isolated mouse liver membranes showed higher titers of aggregation with copper-injured cells than control cells. The in vitro aggregation reaction and blood clearance activity in vivo were abolished by sugars that are known to interact with a hepatic lectin. Our data suggest that copper-induced injury reduces the virulence of Y. enterocolitica and that the liver may be involved in nonimmune rapid clearance of the injured cells, probably by interaction with a hepatic lectin(s).     

Thermal injury of Yersinia enterocolitica.

Procedures were developed to evaluate thermal injury to three strains of Yersinia enterocolitica (serotypes 0:3, 0:8, and 0:17). Serotype 0:17 (atypical strain) was more sensitive to bile salts no. 3 (BS) and to sublethal heat treatment than the typical strains, 0:3 and 0:8. When the 0:3, 0:8, and 0:17 serotypes were thermally stressed in 0.1 M PO4 buffer, pH 7.0, at 47 degrees C for 70, 60, and 12 min, respectively, greater than 99% of the total viable cell population was injured. Injury was determined by the ability of cells to form colonies on brain heart infusion (BHI) agar, but not on Trypticase soy agar (TSA) plus 0.6% BS for serotypes 0:3 and 0:8 and TSA plus 0.16% BS for 0:17. Heat injury of serotype 0:17 cells for 15 min in 0.1 M PO4 buffer caused an approximate 1,000-fold reduction in cell numbers on selective media as compared with cells heated in pork infusion (PI), BHI broth, and 10% nonfat dry milk (NFDM). The extended lag and resuscitation period in BHI broth was 2.5 times greater for 0:17 cells injured in 0.1 M PO4 than for cells injured in BHI or PI. The rate and extent of repair of Y. enterocolitica 0:17 cells in three recovery media were directly related to the heating menstruum used for injury. The use of metabolic inhibitors demonstrated that ribonucleic acid synthesis was required for repair, whereas deoxyribonucleic, cell wall, and protein synthesis were not necessary for recovery of 0:17 cells injured in 0.1 M PO4 buffer, BHI, or PI. Inhibition of respiration by 2,4-dinitrophenol slowed repair only for 0:17 cells injured in 0.1 M PO4 buffer, not for cells injured in PI or BHI.     

Thermal injury of Yersinia enterocolitica

Procedures were developed to evaluate thermal injury to three strains of Yersinia enterocolitica (serotypes 0:3, 0:8, and 0:17). Serotype 0:17 (atypical strain) was more sensitive to bile salts no. 3 (BS) and to sublethal heat treatment than the typical strains, 0:3 and 0:8. When the 0:3, 0:8, and 0:17 serotypes were thermally stressed in 0.1 M PO4 buffer, pH 7.0, at 47 degrees C for 70, 60, and 12 min, respectively, greater than 99% of the total viable cell population was injured. Injury was determined by the ability of cells to form colonies on brain heart infusion (BHI) agar, but not on Trypticase soy agar (TSA) plus 0.6% BS for serotypes 0:3 and 0:8 and TSA plus 0.16% BS for 0:17. Heat injury of serotype 0:17 cells for 15 min in 0.1 M PO4 buffer caused an approximate 1,000-fold reduction in cell numbers on selective media as compared with cells heated in pork infusion (PI), BHI broth, and 10% nonfat dry milk (NFDM). The extended lag and resuscitation period in BHI broth was 2.5 times greater for 0:17 cells injured in 0.1 M PO4 than for cells injured in BHI or PI. The rate and extent of repair of Y. enterocolitica 0:17 cells in three recovery media were directly related to the heating menstruum used for injury. The use of metabolic inhibitors demonstrated that ribonucleic acid synthesis was required for repair, whereas deoxyribonucleic, cell wall, and protein synthesis were not necessary for recovery of 0:17 cells injured in 0.1 M PO4 buffer, BHI, or PI. Inhibition of respiration by 2,4-dinitrophenol slowed repair only for 0:17 cells injured in 0.1 M PO4 buffer, not for cells injured in PI or BHI.     

A rapid and sensitive method for the detection of Yersinia enterocolitica strains from clinical samples

AIMS: To compare three culture methods to detect Yersinia enterocolitica from oral or rectal swabs from experimentally infected pigs. METHODS AND RESULTS: The three methods used were: direct plating on Cefsulodin-Irgasan-Novobiocin (CIN) agar, cold enrichment in phosphate buffered saline (PBS) followed by plating on CIN agar and selective enrichment with Luria-Bertani-Bile Salts Irgasan (LB-BSI) followed by plating on CIN agar. Selective enrichment with LB-BSI produced the highest recovery rate (63%), when compared with cold enrichment (52%) and plating on CIN agar alone (43%). Selective enrichment with LB-BSI was significantly (P < 0.02) more sensitive than direct plating on CIN agar and more sensitive than cold enrichment (P < 0.1). CONCLUSIONS, SIGNIFICANCE AND IMPACT OF THE STUDY: Selective enrichment with LB-BSI was more sensitive than the widely accepted method of cold enrichment and it reduced the time required for detection of Y. enterocolitica by three weeks. Selective enrichment with LB-BSI was also compatible with a multiplex PCR technique.     

Virulence plasmid (pYV)-associated susceptibility of Yersinia enterocolitica to chlorine and heavy metals

This study was aimed at elucidating the role of virulence plasmid (pYV) in the susceptibility of Yersinia enterocolitica to bactericidal agents such as chlorine and heavy metals. Plasmid-bearing (pYV+) Y. enterocolitica was less susceptible to the antimicrobial action of chlorine and heavy metals compared with the isogenic plasmidless (pYV-) derivative. This difference was, however, observed only with bacteria cultured at 25 degrees C. pYV-associated susceptibility apart, cells cultured at 37 degrees C were also found to be less susceptible to the antimicrobial action of these agents. The results indicate that the susceptibility of Y. enterocolitica to these agents was influenced both by the presence of the virulence plasmid and the temperature at which the cells were cultured.     

The effect of temperature and selective agents on the growth of Yersinia enterocolitica serotype O:3 in pure culture

This study examined the individual and combined effects of the selective agents normally present in Yersinia-selective agar (i.e. cefsulodin, irgasan and novobiocin) on the growth kinetics of plasmid-bearing (P+) and plasmid-cured (P-) Yersinia enterocolitica serotype O:3 at 25 and 37 degrees C. Growth studies were carried out in pure culture, and the data obtained were subjected to linear regression analysis to determine lag phase duration(s) and growth rates of the examined strains. In general, the presence of selective agents increased the duration of the lag phase at 37 degrees C, with longer lag phases noted in all cases in which two or more selective agents were present. Growth rates in CIN broth base (CIN NA) and CIN NA plus commercial supplement (SR 109) (CIN) were faster at 37 than 25 degrees C, but in some cultures of incomplete CIN NA broth with less than three supplements added, growth tended to be faster at 25 than 37 degrees C. Generally, plasmid-bearing strains grew slower than plasmid-cured strains in most media at 37 degrees C due to virulence plasmid expression retarding growth. In some instances at 37 degrees C, it was observed that the growth rates of both plasmid-bearing and plasmid-cured strains were comparable, indicating the influence of added selective agent/s negating any effects associated with virulence plasmid expression. The effects of selective agents, incubation temperature and virulence plasmid carriage on the growth kinetics of Y. enterocolitica are discussed.     

Recognition of pathogenic Yersinia enterocolitica by crystal violet binding and polymerase chain reaction

Cefsulodin-Irgasan-Novobiocin (CIN) agar is used for the selective isolation and enumeration of Yersinia enterocolitica from clinical specimens and food. The medium contains crystal violet and about 1 mmol l^-1 calcium and can be used for the phenotypic characterization of strains that carry a virulence plasmid. At 32°C, irrespective of pathogenicity, colonies are translucent with a pale pink centre surrounded by a transparent border ('bullseye'), while at 37°C pathogenic strains grow as calcium-dependent microcolonies which, because of crystal violet binding, are intensely coloured. These results were confirmed by the polymerase chain reaction with primers directed at the vir F gene, which is present only in pathogenic strains of Y. enterocolitica. Pathogenic strains of Y. enterocolitica can be recognized by growth at 37°C on Yersinia selective agar.     

A new chromogenic agar medium for detection of potentially virulent Yersinia enterocolitica

Several outbreaks of foodborne yersiniosis have been documented and this disease continues to be source of infections transmitted through foods. The selective agars most commonly used to isolate Yersinia enterocolitica in clinical, food and environmental samples, cefsulodin-irgasan-novobiocin (CIN) and MacConkey (MAC) agars, lack the ability to differentiate potentially virulent Y. enterocolitica from other Yersinia that may be present as well as some other bacterial spp. This study proposes the use of an agar medium, Y. enterocolitica chromogenic medium (YeCM), for isolation of potentially virulent Y. enterocolitica. This agar contains cellobiose as the fermentable sugar, a chromogenic substrate and selective inhibitors for suppression of colony formation by many competing bacteria. All strains of potentially virulent Yersinia of biotypes 1B, and biotypes 2-5 formed convex, red bulls-eye colonies on YeCM that were very similar to those described for CIN agar. However, Y. enterocolitica biotype 1A and other related Yersinia formed colonies that were purple/blue on YeCM while they formed typical red bulls-eye colonies on CIN agar. When a mixture of potentially virulent Y. enterocolitica biotype 1B, Y. enterocolitica biotype 1A and 5 other bacterial species was used to artificially contaminate tofu and then spread-plated on three selective agars, Y. enterocolitica biotype 1B colonies were easily distinguished from other strains on YeCM. However, Y. enterocolitica biotype 1B colonies were indistinguishable from many other colonies on CIN and only distinguishable from those of C. freundii on MAC. When colonies were picked and identified from these agars, typical colonies from YeCM were confirmed only as Y. enterocolitica biotype 1B. Typical colonies on CIN and MAC were found to belong to several competing species and biotypes.     

Synthesis of a selective agar medium for Yersinia enterocolitica

A new agar medium for isolation of Yersinia enterocolitica was formulated based on growth studies which defined an optimum basal, and the evaluation of selective chemical agents, dyes, and antibiotics. The final formulation, designated cefsulodin-irgasan-novobiocin(CIN) agar, provided quantitative recovery of 40 different strains of Y. enterocolitica in 24 h using incubation at 32 degrees C or with 48 h of incubation at 22 degrees C. The medium was highly selective, especially against Pseudomonas aeruginosa. Escherichia coli, Klebsiella pneumoniae, and Proteus mirabilis. Colony morphology coupled with a differential reaction resulting from mannitol fermentation permitted discrimination of Y. enterocolitica from most of those Gram-negative bacteria that were able to grow on the medium. Recovery and selective characteristics of CIN agar were stable during storage at room temperature for 9 days. CIN agar gave a higher recovery of Y. enterocolitica from feces both direct and with cold enrichment (0.4/1.5%) than Salmonella-Shigella (0.0/0.7%) and MacConkey (0.0/0.9%) agars while significantly reducing the level of background organisms.     

Factors contributing to the reduced invasiveness of chlorine-injured Yersinia enterocolitica.

The invasion of epithelial cells in vitro and in vivo by chlorine-injured Yersinia enterocolitica was assessed by direct microscopic observations. These experiments showed that injury by chlorine inhibited invasiveness of virulent Y. enterocolitica. Two requirements appeared to be necessary for invasiveness: the organism must be viable and metabolically active, and the organism must have certain surface components to initiate engulfment. Inhibition of RNA synthesis by rifampin and protein synthesis by chloramphenicol, tetracycline, and spectinomycin inhibited the invasiveness but not the attachment of Y. enterocolitica to epithelial cells. Membrane preparations from untreated and antimicrobial-agent-treated Y. enterocolitica blocked the invasiveness of virulent Y. enterocolitica, whereas membranes from chlorinated cells were unable to block invasiveness. Chlorine did not change the hydrophobicity or surface charge of injured Y. enterocolitica. The results indicate that invasion was more than simple association of the bacterium with the epithelial cell and involved a specific trigger to stimulate engulfment.     

Factors contributing to the reduced invasiveness of chlorine-injured Yersinia enterocolitica

The invasion of epithelial cells in vitro and in vivo by chlorine-injured Yersinia enterocolitica was assessed by direct microscopic observations. These experiments showed that injury by chlorine inhibited invasiveness of virulent Y. enterocolitica. Two requirements appeared to be necessary for invasiveness: the organism must be viable and metabolically active, and the organism must have certain surface components to initiate engulfment. Inhibition of RNA synthesis by rifampin and protein synthesis by chloramphenicol, tetracycline, and spectinomycin inhibited the invasiveness but not the attachment of Y. enterocolitica to epithelial cells. Membrane preparations from untreated and antimicrobial-agent-treated Y. enterocolitica blocked the invasiveness of virulent Y. enterocolitica, whereas membranes from chlorinated cells were unable to block invasiveness. Chlorine did not change the hydrophobicity or surface charge of injured Y. enterocolitica. The results indicate that invasion was more than simple association of the bacterium with the epithelial cell and involved a specific trigger to stimulate engulfment.     

Membrane damage and microbial inactivation by chlorine in the absence and presence of a chlorine-demanding substrate

The relationship between cell inactivation and membrane damage was studied in two gram-positive organisms, Listeria monocytogenes and Bacillus subtilis, and two gram-negative organisms, Yersinia enterocolitica and Escherichia coli, exposed to chlorine in the absence and presence of 150 ppm of organic matter (Trypticase soy broth). L. monocytogenes and B. subtilis were more resistant to chlorine in distilled water. The addition of small amounts of organic matter to the chlorination medium drastically increased the resistance of both types of microorganisms, but this effect was more marked in Y. enterocolitica and E. coli. In addition, the survival curves for these microorganisms in the presence of organic matter had a prolonged shoulder. Sublethal injury was not detected under most experimental conditions, and only gram-positive cells treated in distilled water showed a relevant degree of injury. The exposure of bacterial cells to chlorine in distilled water caused extensive permeabilization of the cytoplasmic membrane, but the concentrations required were much higher than those needed to inactivate cells. Therefore, there was no relationship between the occurrence of membrane permeabilization and cell death. The addition of organic matter to the treatment medium stabilized the cytoplasmic membrane against permeabilization in both the gram-positive and gram-negative bacteria investigated. Exposure of E. coli cells to the outer membrane-permeabilizing agent EDTA increased their sensitivity to chlorine and caused the shoulders in the survival curves to disappear. Based on these observations, we propose that bacterial envelopes could play a role in cell inactivation by modulating the access of chlorine to the key targets within the cell.     

Evaluation of the usefulness of selected virulence markers for the identification of virulent Yersinia enterocolitica strains. I. Phenotypic markders associated with Plasmid pYV

The species Yersinia enterocolitica includes either pathogenic or non-pathogenic strains. Therefore it is necessary to differentiate virulent bacilli from other. It is well known that pathogenic strains of Y. enterocolitica bearing virulence associated plasmid called pYV, which could be demonstrated by its isolation or detected by the presence of specific, phenotypic properties directly related with this plasmid. The aim of the presented paper was to check the ability of some phenotypic virulence markers associated with pYV, to detection of pathogenic Y. enterocolitica strains. In the presented work 152 (130 carrying pYV) clinical strains of Y. enterocolitica O3 isolated mainly from stool were examined for the presence of phenotypic virulence markers such as: calcium dependency, Congo-red binding, autoagglutination and agglutination with Mangifera indica extract. Both first features were detected parallel, on the same plate, using CRMOX (Congo-red, Magnesium Oxalate) agar. The detection of the tested markers in the examined strains was compared with the presence of virulence plasmid. The obtained results confirmed the observations done by other authors that Y. enterocolitica strains, in which bacilli bearing the virulence plasmid predominate, exhibit all tested phenotypic properties whereas the plasmid-cured isogenic strains show no one of these features. Therefore all the tested markers could be useful for detection of virulent Y. enterocolitica strains directly isolated from patients. The most useful virulence markers in bacteriological study seems to be calcium dependency and Congo-red binding, examined together by the use of CRMOX agar, because they confirm the presence of the virulence plasmid by parallel detection of two physiologically different features associated with this plasmid. In addition CRMOX agar allows for the examination rough strains while agglutination tests do not.     

Luminal ammonia retards restitution of guinea pig injured gastric mucosa in vitro

The present study was conducted to elucidate the mechanisms by which Helicobacter pylori (HP)-derived ammonia causes gastric mucosal injury. Intact sheets of guinea pig gastric fundic mucosae were incubated in Ussing chambers. Both the luminal and the serosal pH were kept at 7.4. Transmucosal potential difference (PD) and electrical resistance (R) were monitored as indices of mucosal integrity. Restitution was evaluated by recovery of PD, R, and transmucosal [(3)H]mannitol flux after Triton X-100-induced mucosal injury. The effects of luminal or serosal NH(4)Cl on function and morphology of uninjured or injured mucosae were examined. In uninjured mucosae, serosal NH(4)Cl induced more profound decreases in PD and R and more prominent vacuolation in gastric epithelial cells than did luminal NH(4)Cl. In contrast, luminal NH(4)Cl markedly inhibited restitution in injured mucosae and caused an extensive vacuolation in gastric epithelial cells, as did serosal NH(4)Cl. Transmucosal ammonia flux was greater in the injured than in the uninjured mucosae. These results suggest that 1) basolateral membrane of gastric epithelial cells is more permeable to ammonia than apical membrane and 2) luminal ammonia, at concentrations detected in HP-infected gastric lumen, retards restitution in injured mucosae.     

Role of RpoS in survival of Yersinia enterocolitica to a variety of environmental stresses.

rpoS, a gene that encodes an alternative sigma factor (also known as katF), is critical for the ability of Yersinia enterocolitica grown at 37 degrees C, but not at 26 degrees C, to survive diverse environmental insults such as high temperature, hydrogen peroxide, osmolarity, and low pH. However, a Y. enterocolitica rpoS mutant was not affected in expression of inv or ail, invasion of tissue culture cells, or virulence in mice.     

USE OF UNIVERSAL PREENRICHMENT MEDIUM SUPPLEMENTED WITH OXYRASETM FOR THE SIMULTANEOUS RECOVERY OF ESCHERICHIA COLI O157:H7 AND YERSINIA ENTEROCOLITICA

Universal Preenrichment (UP) medium was used successfully for the simultaneous recovery of two strains each of Escherichia coli O157:H7 and Yersinia enterocolitica in the presence of Listeria monocytogenes and Salmonella typhimurium. E. coli O157:H7 and Y. enterocolitica populations reached ca. 10⁸ CFU/ml in UP medium in 18 h from an initial level ofca. 10² CFU/ml. Addition of Oxyrase_TM enhanced the growth of both E. coli O157:H7 strains and one strain of Y. enterocolitica. These three strains were able to recover from heat injury by 6 h when 24-h cultures were tested, but not when 18-h cultures were used. Injured and noninjured E. coli O157:H7 could be recovered from artificially inoculated food samples (shredded cheddar cheese, turkey ham, hot dogs, mayonnaise, and ground beef) in UP medium supplemented with Oxyrase^TM (UPO) by 18 h using immunoblotting. Y. enterocolitica could be recovered from turkey ham, hog dogs, and mayonnaise by direct plating on CIN agar from UPO medium. However, recovery of Y. enterocolitica from shredded cheddar cheese and ground beef required subsequent selective enrichment in sorbitol bile broth and isolation on Cefsulodin Irgasan Novobiocin agar (CIN). UPO medium can be used for simultaneous detection of E. coli O157:H7 and Y. enterocolitica from foods. However, subsequent selective enrichment and isolation on selective plating media are required for isolation of Y. enterocolitca from raw foods containing high population levels of background microflora.