Comparison of yoghurt, heat treated yoghurt, milk and lactose effects on plasmid dissemination in gnotobiotic mice.

The effect of yoghurt, heat-treated fermented milk, milk and lactose solution intake on plasmid transfer and establishment of the resulting transconjugants in the digestive tract of mice colonised with human faecal flora were examined. Yoghurt lowered the population level of transconjugants more efficiently than heat-treated fermented milk (-2 log and -1 log respectively) and indicated a beneficial effect of viable bacteria. On the other hand consumption of milk drastically inhibited the establishment of transconjugants, which were below the detection threshold of 10(2) UFC per g of faeces. We were not able to recover transconjugants from faecal samples with lactose supplementation, indicating a possible inhibition of plasmid transfer. Since the yoghurt, heat-treated fermented milk, milk and lactose solution contained approximately the same lactose concentration it is fair to speculate that lactose may contribute to the inhibiting effects of the various supplementations. The inhibitions described were not associated with other intestinal parameters like the intestinal transit time, the population levels of the recipient, or the donor and total anaerobic microflora. It is evident that other parameters need to be investigated such as the composition of the endogenous microflora and metabolic products.

Influence of oral inoculation with plasmid-free human Escherichia coli on the frequency of diarrhea during the first year of life in human newborns.

BACKGROUND: This study was carried out to determine whether early inoculation of the plasmid-free human Escherichia coli into human newborns would reduce the frequency of acute diarrhea during a 1-year period. The plasmid-free E. coli strain isolated from the fecal microbiota of a healthy adult was nontoxigenic in vivo and in vitro and sensitive to all usual antibiotics. METHODS: In the experimental group, 51 healthy newborns were inoculated orally with 106 viable cells of the bacteria within 2 hours after birth. In the control group, the same number of newborns received the heat-killed bacteria. The clinical trial was double blind, and the newborns were randomly assigned to the experimental and control groups. RESULTS: Six months and 1 year after bacterial inoculation, infants in the experimental group showed a higher mean body weight (7.59 +/- 1.15 kg and 9.88 +/- 1.31 kg, respectively; P < 0.05) when compared with the control group (7.03 +/- 1.09 kg and 8.92 +/- 1.38 kg, respectively). At the end of the clinical trial, 48% (23/48) of the infants in the experimental group had shown at least one diarrhea episode during the 1-year period, as opposed to 71% (34/48) in the control group. These values were significantly different (P = 0.037), showing a 32.3% protective effect of inoculation. CONCLUSIONS: The present study shows that protection against diarrhea was obtained by oral inoculation with a single dose of plasmid-free human E. coli soon after birth.

Experimental models of porcine post-weaning colibacillosis and their relationship to post-weaning diarrhoea and digestive disorders as encountered in the field.

The aim of this study was to develop a reliable model system of porcine post-weaning colibacillosis, and in doing so to assess the primary relationship of enterotoxigenic Escherichia coli to post-weaning diarrhoea and digestive disorders as encountered in the field. Six sequential experiments were carried out using 168 SPF piglets weaned into an optimal controlled environment at 28 days of age. The piglets were allocated to 23 treatment groups, 17 of which were inoculated either orally or intragastrically with enterotoxigenic strains of E. coli (LT+, STI+, STII+) possessing adhesive factors including K88 (F4). The piglets were challenged either once (Day 4 post-weaning) or on several days post-weaning, with the challenge load for each inoculation varying from 10(8) to 10(12) CFU. Overall 14.5% of inoculated pigs developed severe illness and died: these had lesions in their digestive tracts typical of colibacillosis. Diarrhoea occurred on at least 1 day in 50% of inoculated pigs, but was transient (1.7 days on average), appeared very soon after challenge (sometimes within half a day), and was accompanied by signs of depression and low weight gain. Generally a prompt recovery then occurred. In the second 2 weeks post-inoculation daily weight gain reached the same level in most inoculated groups of pigs as in the uninoculated controls. Only a small number of pigs developed a chronic enteritis lasting several days, as is typically observed in field cases. Diarrhoea was more common in the piglets that were tested adhesive positive to the K88 fimbriae receptor, but the disorders were no more severe in these animals. The response of all pigs depended primarily on the inoculum used, and especially on the challenge load. Although enterotoxigenic E. coli are clearly important in the aetiology of post-weaning diarrhoea, other factors are also required for the production of the chronic post-weaning digestive disorders and ill-thrift that are commonly encountered in commercial piggeries.

Effects of yoghurt intake on plasmid transfer and colonisation with transconjugants in the digestive tract of mice associated with human faecal flora.

This study deals with the effects of yoghurt intake on wild-type and recombinant plasmid transfer from an exogenous Escherichia coli K12-derivative donor strain to an endogenous recipient strain in the digestive tract of mice associated with human faecal flora. We showed that the self-transmissible plasmid R388 was efficiently transferred to recipient strain PG1 in mice associated with human faecal flora (HFF-PG1) and that the resulting transconjugants (PG1-R388) became established at a high and maximal population level without any selective pressure. Using HFF-PG1 mice made it possible to determine whether yoghurt consumption decreases R388 transfer efficiency and the ability of transconjugant PG1-R388 to successfully colonise the digestive tract. Results indicated that yoghurt consumption had two effects: it reduced the efficacy of plasmid transfer 10-fold and decreased the transconjugant PG1-R388 population density 100-fold, compared to the control group. We also describe another HFF mouse model in which recipient PG1 was replaced by EM0 with which no plasmid transfer was observed. This model made it possible to demonstrate the potential promoting effect of yoghurt intake on transconjugant formation and establishment. Our results indicated no yoghurt effect; no transconjugants appeared in the digestive tract of HFF-EM0 mice fed on yoghurt or on standard food. In both mouse models, HFF-PG1 and HFF-EM0, yoghurt intake did not promote the mobilisation of either the non-self-transmissible plasmid pUB2380 or the recombinant plasmid pCE325.

Effects of racecadotril and loperamide on bacterial proliferation and on the central nervous system of the newborn gnotobiotic piglet.

METHODS: The effects of 4 days of oral administration of different doses of two drugs, an enkephalinase inhibitor (the antisecretory agent, racecadotril) and a mu-receptor agonist (loperamide), on intestinal growth of a bacterial nonpathogenic strain (Escherichia coli E 404) and on the central nervous system (CNS) were compared in newborn gnotobiotic piglets. RESULTS: The E. coli content of the proximal jejunum (segment S1) and the E. coli ratio of stomach:segment S1 were similar in the racecadotril (20 mg/kg b.d., n = 5) and control groups. In contrast, in the loperamide group (1 mg/kg b.d., n = 4), the E. coli content of segment S1 and the E. coli ratio stomach:S1 were both significantly higher than with racecadotril or control (P = 0.04 and 0.005, respectively, for E. coli content; P = 0.05 and 0.03, respectively, for stomach:S1). There were no clinical signs of neurotoxicity and no deaths with racecadotril given orally at a high dose of 130 mg/kg b.d. (n = 5)--nearly 60 times the paediatric dosage. In contrast, an equivalent high dose of loperamide (5 mg/kg b.d.) resulted in death in three out of four piglets. CONCLUSIONS: In contrast to loperamide, racecadotril did not induce bacterial overgrowth and did not produce central neurotoxicity.

Effect of fermented milk intake on plasmid transfer and on the persistence of transconjugants in the digestive tract of gnotobiotic mice.

Plasmid transfer occurs in the digestive tract and the transconjugants may become durably established. The aim of the present work is to investigate the effect of probiotics of plasmid transfer and on establishment of transconjugants in the gut. Plasmid transfers were carried out in the digestive tract of germ free mice associated with an E. coli K12 donor strain harboring three plasmids (R388, self-transmissible, pCE325 and pUB2380, mobilisable,) and an E. coli recipient strain, PG1, of human origin (Duval-Iflah et al., 1994). Milks fermented with either Lactobacillus bulgaricus or Streptococcus thermophilus or symbiosis, S85, of both strains were given daily as 1/3 of food diet. Fermented milks have no effect on the transfer of R388 and pUB2380 except a slight increase of TC(R388) with milk fermented with S85. Long term ingestion of milk fermented with S85 inhibited the formation and the establishment of transconjugants TC(pCE325). Milk fermented with L. bulgaricus lowered the population density of TC(pCE325) in animals where they were already established. This phenomenon was reversible, since the density of TC(pCE325) increased in the same animals after cessation of supplementation. Bacterial cultures obtained in MRS broth and given in state of drinking water were compared with fermented milks. Bacterial cultures with L. bulgaricus and with S85 favoured the establishment of TC(pCE325). There results indicate for the first time that probiotics have various effects on the formation and/or establishment of transconjugants in the gut of axenic mice. The effects depend on whether the probiotics were cultivated in milk or in MRS, indicating that bacterial metabolites and viable bacteria can be involved.

Recombinant plasmid mobilization between E. coli strains in seven sterile microcosms.

Transfer by mobilization of a pBR derivative recombinant plasmid lacking transfer functions (oriT+, tra-, mob-) from one E. coli K12 strain to another was investigated in seven sterile microcosms corresponding to different environments. These microcosms were chosen as representative of environments that genetically engineered microorganisms (GEMOs) encounter after accidental release, namely attached biomass in aquatic environments (biofilm), soil, seawater, freshwater, wastewater, mouse gut, and mussel gut, GEMOs survived in the same way as the host strains in all microcosms. Recombinant DNA mobilization occurred in the mouse gut, in sterile soil, and in biofilm. The plasmid transfer rates principally reflected the environmental conditions encountered in each microcosm.

The porcine intestinal receptor for Escherichia coli K88ab, K88ac: regional localization on chromosome 13 and influence of IgG response to the K88 antigen.

The loci encoding the porcine intestinal receptors for Escherichia coli K88ab and K88ac (K88abR and K88acR) were firmly assigned to chromosome 13 by linkage analysis using a three-generation pedigree. The linear order of these loci and seven other markers on chromosome 13 was determined by multipoint analyses. The K88abR and K88acR loci were tightly linked (theta = 0.01, zeta = 41.06) with the K88abR locus localized 7.4 cM (sex average) proximal to the transferrin locus. The results, together with previous reports from two other groups, provide an unequivocal assignment of the K88 receptor loci to chromosome 13, and reject a previous assignment to chromosome 4. Pigs possessing the receptor had a slightly higher specific IgG response to the K88 antigen after an intramuscular immunization with an E. coli vaccine.

Use of porcine fibrinogen as a model glycoprotein to study the binding specificity of the three variants of K88 lectin.

Known glycoproteins were used to determine the differences occurring in the binding specificities of the three variants of the K88 lectin in an approach essentially based on lectin blotting. During the screening, it was demonstrated that each variant of the K88 lectin biotinylated via its amino groups (NbioK88) exhibited a characteristic binding to the three chains of porcine fibrinogen. NbioK88ab weakly bound to A alpha chains, NbioK88ac bound to B beta and gamma chains, and NbioK88ad bound only to the gamma chain. To validate this model, the oligosaccharide moieties of porcine fibrinogen were analyzed with glycosidases and by lectin blotting and sugar composition. Both the B beta chain and gamma chain carry biantennary N-glycans of the N-acetyllactosamine type that are not recognized by K88 lectins. A alpha chains are substituted by sialylated T antigen. O-glycans were also detected on B beta and gamma chains of porcine fibrinogen and contribute to the recognition of these chains by K88ac and K88ad fimbriae.

Characterization of O-glycan moieties of the 210 and 240 kDa pig intestinal receptors for Escherichia coli K88ac fimbriae.

The porcine brush border glycoproteins of 210 and 240 kDa, recognized by Escherichia coli K88ac fimbriae, contained O-linked oligosaccharides. The carbohydrate moieties were analysed by deglycosylation, lectin-binding and agglutination assays. Neuraminidase susceptibility of the 210 kDa receptor suggested that a sialoglycoprotein may act as receptor for the K88ac fimbriae. In contrast, K88ac-binding to the 210 and 240 kDa glycoproteins totally disappeared after fucosidase treatment, indicating the critical role of fucosyl residues at the receptor sites. Among the oligosaccharides extracted from these O-glycoproteins, K88ac fimbriae showed affinity for neutral sugar chains while sialylated species were not recognized. Our data suggest a possible role of the polypeptide backbone in the definition of receptor sites. Specific agglutination by K88ac-fimbriated E. coli of the erythrocytes of the hamster Mesocricetus auratus was inhibited by the anti-T peanut lectin and the lectins of Datura stramonium, Aleuria aurantia and Maackia amurensis. Hence, we propose that Gal beta 1-3GalNAc- and Fuc alpha 1-2Gal beta 1-3/4GlcNAc- are the main sequences mediating K88ac fimbrial binding. These structures were not detected in the non-adhesive piglet brush borders characterized by a high carbohydrate content. Additional oligosaccharides probably masked the underlying receptor structures.

Study of gene transfer in vitro and in the digestive tract of gnotobiotic mice from Lactococcus lactis strains to various strains belonging to human intestinal flora.

The use of genetically modified organisms (GMO) in dairy products requires evaluation of the DNA transfer capacity from such organisms among the human intestinal microflora. Thus, both in vitro and in vivo [in the digestive tract (DT) of mice] transfer from Lactococcus lactis donor strains of the conjugative plasmid pIL205 (CmR) and the non-conjugative plasmid pIL253 (EmR) to: (1) recipient strains isolated from human faecal flora Bacteroides sp., Bifidobacterium sp., Peptostreptococcus sp. (strictly anaerobic bacterial strains) and Enterococcus faecalis, (2) a whole human faecal flora, was studied. In both cases, no gene transfer was observed to strictly anaerobic bacterial strains. DNA transfer was only observed to the E. faecalis strain: in vivo CmR E. faecalis transconjugants were isolated from sequentially multi-associated mice and when the recipient strains associated with the mice, they were a defined mixture of Bacteroides sp., Bifidobacterium sp., Peptostreptococcus sp. and E. faecalis strains. When mice were associated with the whole human faecal flora, the plasmid pIL205 was transferred into some facultative anaerobic streptococci. It was also shown that DNA transfer occurred even when the lactococcal donor strain was transient in the DT of the gnotobiotic host animals.

Gene transfer from engineered Lactococcus lactis strains to Enterococcus faecalis in the digestive tract of gnotobiotic mice.

The introduction of genetically modified organisms into food products requires an evaluation of the behaviour and the dissemination of foreign genes of such organisms among the human intestinal microflora. The conjugal transfer, both in vitro and in vivo (in mice digestive tract) of DNA from Lactococcus lactis donor strains to an Enterococcus faecalis strain isolated from human faecal flora was studied. We followed the transfer of (1) the self-transmissible plasmid pIL205; (2) two non-self-transmissible but mobilizable plasmids, pIL252 and pIL253; (3) one plasmid, pMS1.5B, integrated into the chromosome of L. lactis. In vitro, the transfer frequency of pIL205 (expressed as the number of transconjugants per donor cell) was 9.6 x 10(-4); mobilization of one of the non-self-transmissible plasmids, pIL253, was observed (4.9 x 10(-7)). In vivo, only transfer of pIL205 and pIL253 occurred, but the frequency was not determined. The transfer of pMS1.5B was not detected in vitro or in vivo.

Evidence for linkage between K88ab, K88ac intestinal receptors to Escherichia coli and transferrin loci in pigs.

Segregation at the loci coding for the K88ab and K88ac small intestinal receptors to E. coli adhesins (K88abR, K88acR) and at the transferrin (TF) locus was studied in 38 pig families including 273 piglets. The TF locus showed a segregation deviation towards the B variant while each of the K88 receptors behaved as a single autosomal dominant gene. Recombinants between K88abR and K88acR provide evidence that they are under the control of two different loci. Thirty-two triple backcross families were selected to test linkage and estimate recombination rates (theta). Our results demonstrate that the two K88 receptor loci are closely linked (theta = 0.02) with a maximum lod score value (Zm) of 46.0. In addition, they are linked to the TF locus, theta = 0.14, Zm = 19.6 for the K88abR locus and theta = 0.16, Zm = 17.9 for the K88acR locus. The estimated recombination rates, smaller in males than in females, are consistent with the order TF-K88abR-K88acR. This linkage thus localizes the K88 loci, as the TF locus, on chromosome 13.

Colonization of the digestive tract of germ-free mice by genetically engineered strains of Lactococcus lactis: study of recombinant DNA stability.

The ability of genetically engineered Lactococcus lactis strains to become established in the digestive tract (DT) of germ-free mice was examined together with the stability of their genetic markers. Seven L. lactis strains were genetically modified by insertion of genetic markers on different replicons: chloramphenicol resistance gene cat was carried by self-transmissible plasmid pIL205, a derivative of plasmid pIP501; erythromycin resistance gene erm, originating from pAM beta 1, was inserted into non-transmissible plasmids pIL252 and pIL253 of low and high copy number respectively; erm gene from plasmid pMS1.5B was inserted into the chromosome. All strains carried a common wild-type plasmid pIL9 involved in lactose fermentation. It was observed that the DT of mice was rapidly and efficiently colonized with either the inoculated parental strain or with its derivatives or with both of them, but plasmid-free derivatives were always at dominant levels. Both plasmids pIL9 and pIL205 were lost, but the parental strains and the plasmid-lacking derivatives were at codominant levels, indicating that there is an equilibrium between plasmid loss and plasmid transfer in the DT. Strains that carried non-transmissible and low copy number plasmid pIL252 were rapidly eliminated from the DT, which in turn was colonized with the respective pIL252-less derivatives; this is probably due to the high segregational instability of pIL252.(ABSTRACT TRUNCATED AT 250 WORDS)

Adhesion of K99 fimbriated Escherichia coli to pig intestinal epithelium: correlation of adhesive and non-adhesive phenotypes with the sialoglycolipid content.

Evidence for the existence of two phenotypes of piglets born to experimental herds was obtained based on the susceptibility of intestinal brush borders to adhesion of K99-positive Escherichia coli. The enterocytes of the K99-receptive piglets displayed a characteristic sialoglycolipid pattern, with a higher content of the monosialoglycolipids II3NeuGc-LacCer (GM3Gc), IV3NeuGc-nLcOse4Cer (SPGGc) and IV3NeuAc-nLcOse4Cer (SPG) and the oligosialogangliosides IV3NeuAc,II3NeuAc-GgOse4Cer (GD1a), II3(NeuAc)2-GgOse3Cer (GD2), II3(NeuAc)2-GgOse4Cer (GD1b) and IV3NeuAc,II3(NeuAc)2-GgOse4Cer (GT1b) when compared to the gangliosides of non-receptive piglets. The gangliosides from enterocytes of the non-receptive piglets were mainly the monosialogangliosides II3NeuAc-GgOse3Cer (GM2) and II3NeuAc-LacCer (GM3), only traces of the other sialoglycolipids being detected. Adhesion of 14C-labelled K99-positive E. coli cells to the piglet small intestinal sialoglycolipids, as tested by the thin-layer chromatogram overlay assay, revealed that the receptive enterocyte membrane was richer in glycolipids containing K99 receptor structures than the non-receptive enterocyte. Adhesion of K99-positive E. coli correlated with the degree of sialylation of the brush border glycolipids.

Double-sandwich enzyme-linked immunosorbent assay for determination of Escherichia coli heat-labile porcine enterotoxins.

A "double-sandwich" ELISA for the detection and measurement of a heat-labile enterotoxin produced by porcine enterotoxigenic strains of Escherichia coli (LTp) is described. In contrast with other heat-labile toxins, LTp did not bind to agarose gels and exhibited a very low affinity for GM1 in the classical GM1-ELISA technique. The similarity of LTp with cholera toxin was confirmed by immunoblotting. This property allowed the binding of LTp to rabbit IgG anti-cholera toxin antibodies (covalently linked to polystyrene plates) and sheep anti-cholera toxin serum. The immunocomplex was revealed by anti-sheep immunoglobulin antibodies conjugated with peroxidase. Application of the "double-sandwich" ELISA to the quantitation of toxin production by two strains, which differ only in the presence or the absence of the K88ab antigen, showed that the Ent+, K88+ strain produced significantly less toxin than the Ent+, K88- derivative.

Resistance of gnotobiotic large white and Chinese piglets to in vivo attachment of a K88ab enterotoxigenic Escherichia coli strain.

In vivo adhesion of a K88ab-positive enterotoxigenic Escherichia coli (ETEC) strain to the small intestinal wall of gnotobiotic, colostrum-deprived Chinese and Large White piglets was investigated. A non-enterotoxigenic, attachment factor-deprived E. coli strain, inoculated in association with the K88ab ETEC strain, was used as a marker to determine the content residues on the intestinal walls of gnotobiotic piglets. Both strains were selectively numerated in the luminal contents and on the washed wall from three segments of the small intestine. In vivo attachment was assessed by the ratio between the number of K88ab ETEC adherent to the wall and the number of both the E. coli strains which came from the luminal content residues and were not completely removed by washing. This ratio was first calculated in one group of 8 Large White piglets associated with the marker strain and with a K88ab antigen-deprived E. coli which derived from the parental K88ab ETEC strain. Thus, the range of the ratio was determined in those piglets where no attachment was expected. A comparison between the latter values of the ratio and the values obtained from 15 Large White and 10 Chinese piglets inoculated with the marker strain and the K88ab ETEC strain allowed us to classify the Large White piglets into adhesive (8 piglets) and non-adhesive (7 piglets) phenotypes and to show that the 10 Chinese piglets belonged to the non-adhesive phenotype.

Proteins in the digesta of the pig: amino acid composition of endogenous, bacterial and fecal fractions.

When studying digestibility, the respective parts of the exogenous, endogenous and bacterial fractions in the digesta or feces must be measured. The proportions of proteins from different sources may be estimated by comparing their amino acid composition with those of reference sources. This study describes the composition of endogenous and microbial proteins, i.e. meconium of piglet small intestine and colon, axenic piglet feces, bacteria isolated in the feces of pigs receiving a standard (cereal-based) or a purified diet, and pure culture of Escherichia coli. The composition of monoxenic piglet feces and of feces of conventional pigs fed the two types of diets have also been studied. The data on 17 amino acids were used to make overall comparisons of compositions, using the method of X2 distances and factorial correspondance analysis. The composition of exclusively endogenous products differed somewhat from that of samples (mucus, mucosa) usually considered as representative of that fraction. In conventional pigs, the major part of fecal proteins was composed of bacterial protein. Accurate estimation of these was difficult: diaminopimelic acid assay gave an estimate of 65% bacterial protein, while in the same experimental conditions X2 distance gave an estimate of 90%.

[Effect of orally administered bovine lactoferrin and bovine IgG on the establishment of Escherichia coli in the digestive tract of gnotobiotic mice and human newborn infants]. / Effet de la lactoferrine bovine et des IgG bovines données per os sur l'implantation de Escherichia coli dans le tube digestif de souris gnotoxéniques et de nouveau-nés humains.

The bacteriostatic effect of the association bovine lactoferrin (LF) and bovine IgG (IgG) was studied in vitro and in vivo against two Escherichia coli strains, S17 and EMO1, isolated from the faecal flora of mouse and man, respectively. These two strains were sensitive in vitro to the bacteriostatic effect of LF + IgG. A kinetic study of the in vivo establishment of E. coli S17 was followed in axenic mice associated with that strain. Seven hours after inoculation, no difference was observed in the faecal level of E. coli between control mice and mice fed the same diet supplemented with LF + IgG. An in vivo study was also carried out in human newborns receiving either maternized milk (Nursie) or the same milk supplemented with LF + IgG during the first 48 h of life. One group of babies was inoculated at birth with E. coli EMO1, while another was not. Between the ages of 1 and 5 days, the kinetics of establishment of the E. coli strains spontaneously found in the digestive tract of non-inoculated babies was not significantly different between the group which received milk supplemented with LF + IgG and that which did not. This result was confirmed in infants inoculated with E. coli EMO1. Likewise, the faecal levels of E. coli EMO1 were similar in the supplemented and non-supplemented babies, and, already from day 1, the population level was high and only slight individual variations between babies of the same group were observed. These findings show that the in vitro bacteriostatic effect of LF + IgG on the growth of E. coli strains is not found in vivo.