Pathogenicity of porcine enterotoxigenic Escherichia coli that do not express K88, K99, F41, or 987P adhesins.

Three-week-old weaned and colostrum-deprived neonatal (less than 1 day old) pigs were inoculated to determine the pathogenicity of 2 enterotoxigenic Escherichia coli isolates that do not express K88, K99, F41, or 987P adhesins (strains 2134 and 2171). Strains 2134 and 2171 were isolated from pigs that had diarrhea after weaning attributable to enterotoxigenic E coli infection. We found that both strains of E coli adhered in the ileum and caused diarrhea in pigs of both age groups. In control experiments, adherent bacteria were not seen in the ileum of pigs less than 1 day old or 3 weeks old that were noninoculated or inoculated with a nonpathogenic strain of E coli. These control pigs did not develop diarrhea. Antisera raised against strains 2134 and 2171 and absorbed with the autologous strain, grown at 18 C, were used for bacterial-agglutination and colony-immunoblot assays. Both absorbed antisera reacted with strains 2134 and 2171, but not with strains that express K99, F41, or 987P adhesins. A cross-reaction was observed with 2 wild-type K88 strains, but not with a K12 strain that expresses K88 pili. Indirect immunofluorescence with these absorbed antisera revealed adherent bacteria in frozen sections of ileum from pigs infected with either strain. We concluded that these strains are pathogenic and express a common surface antigen that may be a novel adhesin in E coli strains that cause diarrhea in weaned pigs.     

Comparison of seroagglutination, ELISA, and indirect fluorescent antibody staining for the detection of K99, K88, and 987P pilus antigens of Escherichia coli.

Seroagglutination (SAT), enzyme-linked immunosorbent assay (ELISA), and indirect fluorescent antibody staining tests (IFAT) were compared for reliability in the detection of pilus antigens K99, K88, and 987P of Escherichia coli. Test sensitivities were compared using mixtures of piliated bacteria of several strains diluted to a constant optical density with a nonpiliated strain. Relative sensitivities and specificities of the 3 tests were also compared using 55 E. coli strains that had previously been serotyped and characterized for pilus genes by DNA probe. Although specificity was not a serious problem with any of the tests, the SAT was relatively nonsensitive. The IFAT showed the greatest sensitivity of the 3 tests in detecting K88, K99, and 987P E. coli.     

Use of immunofluorescence, Gram's staining, histologic examination, and seroagglutination in the diagnosis of porcine colibacillosis

An indirect fluorescent antibody (IFA) test of ileal impression smears for K88, K99, and 987P pilus antigens was compared with histologic examination and seroagglutination of Escherichia coli isolates for efficacy in determining colibacillosis in pigs. Histologic examination appeared to be more effective than the IFA test in revealing colonization of the ileum by bacteria. However, histologic examination revealed little about the nature of the colonizing bacteria. Correlation between bacterial adherence, as observed in histologic sections of ileum, and the presence of bacteria with adherence pili, as determined by IFA testing, was 91%. Results of seroagglutination for pilus antigens correlated with results of histologic examination in only 84% of the cases. Pilus antigens were not identified by IFA testing or seroagglutination in 5% of the cases in which adherent bacteria were observed in histologic sections of the small intestine. Because little preparation time was required for the IFA test, results were available within 2 hours of necropsy. In contrast to histologic examination, the IFA test made possible identification of the colonizing organism as E coli and revealed the type of pilus antigen present. Adherence pili were identified more frequently by IFA testing than by seroagglutination. Examination of Gram's-stained ileal impression smears was useful in screening for colibacillosis in pigs. Bacterial adherence was found in positive correlation with the number of gram-negative bacilli observed. Bacterial adherence rarely was observed in histologic sections of ileum when smears contained less than or equal to 10 gram-negative bacilli/1,000 X microscopic field. Each diagnostic test compared offered advantages and disadvantages over the other tests. Seemingly, concurrent use of several of these tests, rather than one, should be used in the diagnosis of porcine colibacillosis.     

Iscom (immunostimulating complex) vaccines containing mono- or polyvalent pili of enterotoxigenic E. coli; immune response of rabbit and swine

Iscom (immunostimulating complex) vaccines were prepared to contain K88ab, K88ac, K99 and 987P pili (fimbriae) of enterotoxigenic E. coli bacteria as monovalent or quadrivalent preparations. The iscoms injected into rabbits and into pigs elicited similar or higher immune response in both animal species than the oil adjuvanted vaccine containing about 5 times more of the same pilus protein. It is concluded that inclusion of pili into iscoms results in immunogenic preparations likely worth pursuing for vaccine production against enterotoxic colibacillosis of newborn pigs. The iscoms did not induce local reaction at the injection sites in contrast to the oil adjuvanted vaccines.     

Fimbriae and enterotoxins associated with Escherichia coli serogroups isolated from pigs with colibacillosis

A comprehensive study of 223 Escherichia coli isolates from pigs with colibacillosis included determination of O serogroups, detection of heat-labile enterotoxin, heat-stable enterotoxin (STa and STb), and identification of K88, K99, 987-P, F-41, and type 1 fimbriae. The incidence of the various E coli types among isolates of pigs of different ages was also determined. Escherichia coli bearing K88 fimbriae accounted for 48% of all isolates studied, were most often of serogroup O157, O149, or O8, and usually produced labile toxin alone or in combination with STa or STb. These E coli were commonly isolated from pigs in each age group studied (0 to 5 days, 6 to 10 days, 11 to 24 days, and greater than 24 days). Escherichia coli bearing 987-P accounted for 30% of the isolates, were most often of serogroup O141 or O20, and usually produced STa. Escherichia coli bearing K99 accounted for 13% of the isolates, usually were of serogroup O101 or O8, and almost always produced STa. Escherichia coli bearing 987-P or K99 were most often isolated from pigs less than 6 days of age. Fimbriae F-41, when identified, were usually on E coli of serotype O101:K99. Although infrequently found, type 1 fimbriae were on E coli of most of the serogroups identified in this study.     

Susceptibility of Chinese Meishan and European large white pigs to enterotoxigenic Escherichia coli strains bearing colonization factor K88, 987P, K99, or F41

Conventionally raised Chinese Meishan and European Large White pigs were intragastrically challenge exposed with 2.1 x 10(10) enterotoxigenic Escherichia coli strains bearing colonization factor K88, 987P, F41, or F41 plus K99. In response to challenge exposure with the K88-positive (K88+) organisms, 96% of Large White pigs died within 48 hours, whereas none of the Meishan pigs died. Both breeds of pigs had similar susceptibility to strains bearing 987P or F41. Lastly, Meishan pigs were found to be more susceptible than Large White pigs to a strain expressing K99 and F41. In pigs with diarrhea, challenge-exposure strains intensively colonized the jejunum (10(8) to 10(10) bacteria/g of tissue) and, to less extent, the duodenum (except K88+ strain, which comprised 10(8)/g). In most cases, jejunal concentrations of the challenge-exposure strains were substantially lower in pigs that did not have diarrhea. Half the resistant Meishan pigs eliminated the K88+ strain from the intestines. Colostral antibody titer that agglutinated challenge-exposure strains did not differ between Meishan and Large White gilts. Results indicate that resistance of pigs to the K88+ strain did not extend to enterotoxigenic strains bearing other well-known factors. They indicate, in addition, that genetic resistance to K88+ strains described in pigs in Europe may exist in pigs in China.     

Fimbriae in Escherichia coli isolated from the small intestine of piglets

Ninety E. coli strains, isolated from piglets which had died from neonatal diarrhea, were tested for the presence of K88, K99, 987P and type 1 fimbriae. Two or more types of fimbriae were demonstrated in 14 of the strains, a single fimbria! type in 44 strains while in 32 strains no fimbriae were detected. Of the 14 E. coli strains with more than 1 type of fimbriae, 12;, 10, 8 and 4 strains showed K88, K99, 987P and type 1, respectively.Twelve E. coli strains were isolated from piglets which had died in the neonatal period without showing signs of neonatal diarrhea at necropsy. One strain showed 987P and 3 strains showed type 1 fimbriae, while the remaining 8 strains were unfimbriated.Sixteen fimbriated E. coli strains were subcultured in order to examine the stability of fimbrial expression in the strains. The K88 and the type 1 fimbriae were regularly expressed, while the K99 and 987P were inconsistently demonstrated.     

Prevalence of four enterotoxin (STaP, STaH, STb, and LT) and four adhesin subunit (K99, K88, 987P, and F41) genes among Escherichia coli isolates from cattle

Colony hybridizations with DNA probes for 3 heat-stable (STaP, STaH, and STb) enterotoxins and 1 heat-labile (LT) enterotoxin and for 4 adhesins (K99, F41, K88, 987P) were performed on 870 Escherichia coli isolates to determine pathotypes prevalent among enterotoxigenic E coli (ETEC) isolated from cattle in Belgium. One hundred thirty-two E coli isolates (15.2%) hybridized with probes STaP, K99, and/or F41. The 5 other probes were not hybridized by E coli isolates. Therefore, only STaP enterotoxin and K99 and F41 adhesins were virulence factors of ETEC isolated from cattle. Two major pathotypes accounted for 95% of the ETEC: STaP+K99+F41+ (67.4%) and STaP+K99+ (27.3%). The last 5% of probe-positive isolates had STaP+, STaP+F41+, or K99+F41+ minor pathotypes. Of 12 American ETEC isolates also assayed, 7 were positive with STb and/or 987P probes (pathotypes STaP+STb+, STaP+ 987P+, or STaP+STb+987P+) and may be porcine- rather than bovine-specific enteropathogens. The remaining 5 American ETEC isolates belonged to 3 minor pathotypes (STaP+, STaP+F41+, and K99+F41+) also found among Belgian E coli isolates. Such isolates may be derivatives of STaP+K99+F41+ or STaP+K99+ ETEC after in vivo or in vitro loss of virulence genes and/or non-ETEC isolates, which have acquired virulence genes by in vivo transfer.     

Development of vaccines for bacterial diseases using recombinant DNA technology

A vaccine was prepared using recombinant DNA techniques to prevent fatal enterotoxigenic Escherichia coli diarrhea in swine. The product, which is a subunit vaccine, was prepared by mechanical and chemical removal of pilus adhesins from the surface of genetically engineered strains of E. coli. The vaccine contains the pilus adhesins K88, K99, and 987P plus an adjuvant. The genes responsible for production of K88 and K99 were separately cloned into the multicopy vector pBR322. K88 was found to be encoded on a 7.6-kilobase HindIII-EcoRI fragment, and K99 was found to be encoded on a 7.15-kilobase BamHI fragment. Strains containing the recombinant plasmid for K99 produced up to ten times more K99 than strains containing the wild-type plasmid. Vaccination of pregnant pigs with the vaccine led to production of pilus-adhesin-specific antibodies that were transferred to the piglets in colostrum and milk. Pilus-adhesin-specific antibodies neutralized the adhesiveness of the pili on enterotoxigenic E. coli, thus preventing attachment, colonization, and disease. Mortality of pigs in litters from vaccinated pigs due to experimentally induced enterotoxigenic E. coli diarrhea was reduced 10-to-20-fold (depending upon the challenge strain), and the incidence, severity, and duration of diarrhea were also reduced.     

Evaluation of a monoclonal antibody to the K99 fimbrial adhesin produced by Escherichia coli enterotoxigenic for calves, lambs and piglets

All the K99+ Escherichia coli grown at 37 degrees C stained strongly with a peroxidase labelled K99 monoclonal antibody using a direct immunoperoxidase staining procedure. There was no reaction when these bacteria were cultured at 18 degrees C or when K99- E coli were grown at either temperature. The binding of the monoclonal antibody to K99 antigen was inhibited by OK antisera to heterologous K99+ E coli but OK antisera to E coli producing adhesins other than K99 were without effect. Using the slide agglutination test the reactions of the monoclonal antibody were identical to those of a polyclonal antiserum to K99 when both were used in parallel to examine 100 K99+ E coli from at least 10 somatic O groups and 1308 K99+ E coli from at least 82 different somatic O groups submitted for routine serological typing in England or the, USA. The monoclonal antibody reacted with K99+ E coli in cryostat sections of the ileum from a piglet infected with E coli strain B44 (O9: K30, K99, F41) but there was no reaction with similar material from piglets infected by E coli strains 1751 (O101: F41), X177/81 (O9: K103, 987P) or Abbotstown (O149: K91, K88ac).     

The use of the fluorescent antibody test to detect Escherichia coli with K88 and 987P pilus antigens

Direct fluorescent antibody tests were used to detect Escherichia coli possessing K88 and 987P antigens. Identification of bacteria was accomplished on suspensions of organisms from clinical isolates, on frozen sections and impression smears from small intestine and on faecal smears. This assay makes possible the rapid identification of E. coli possessing K88 and 987P pilus antigens.     

Antibody response in mice, rabbits and pigs in response to vaccination with an inactivated oil vaccine containing rotavirus and Escherichia coli strains with K88, K99 and 987P fimbrial antigens]

In trials with mice, rabbits and weanling piglets, four experimental charges of a combined inactivated oil vaccine against diarrhoeas in mammals were tested: the vaccine was to be implanted to sows and it contained porcine rotavirus (PRV); two charges also contained bovine rotavirus and bacterins of enterotoxicogenic strains of E. coli with protective antigens K88, K99 and 987P. At low starting antibody titres the twofold i.m. implantation of 0.2 ml vaccine stimulated in mice the production of antibodies to reach the average titre value of 1:128 against PRV and of 1:256 against BRV; in rabbits the twofold i.m. implantation of 2 ml vaccine stimulated the antibody development to reach the average titres of 1:508 or 1:500, and in weanlings after the twofold i. m. implantation of the vaccine the titres were 1:1028 or 1:469; in mice agglutination antibodies to antigen K88 had the average value of 1:68, to antigen K99 the value of 1:44 and to antigen 987P the value of 1:8192; in rabbits the respective titres were 1:285, 1:136 and 1:6006 and in pigs 1:570, 1:631 and 1:8192. The antibodies to antigen 987P persisted at the same level in pigs for six months. Even though there was a gradual decrease in the antibodies to antigens K88 and K99, at that time the values were 9.8 times, or 15.2 times higher than the starting values, and only the antibodies to PRV dropped to the pre-vaccination level. Repeated administration of vaccine to pigs after six months from revaccination induced, with the exception of antigen 987P, an increase in antibodies in a fortnight to reach such titres that were recorded after revaccination.     

Preliminary characterization of Escherichia coli isolated from pigs with diarrhoea in Venezuela

The aetiology of neonatal porcine diarrhoea was studied in 15 different herds located in the north-western region of Venezuela. Of 56 strains of Escherichia coli analyzed, 16 (28.6%) were shown to produce heat-stable (STa) enterotoxin, as detected by infant mouse assay. Only four of these STa+ isolates also possessed the K88 pilus antigen, two were 987P+ and none possessed the K99 antigen, leaving 10 STa+ samples in which no pilus antigen was identified. Among the 40 STa negative samples were six K88+ specimens, one K99+, four 987P+, one which reacted as K88+ + K99+ and one K88+ + 987P+. Considering as pathogenic any strain showing at least one of the characters studied, pathogenic E. coli were detected with an overall frequency of 42.9%, being more prevalent during the second week of life. An electrophoretic analysis of the plasmid content of the field isolates of E. coli, revealed the presence of numerous species of extrachromosomal DNA, although no direction association could be made between a particular plasmid and any of the pathogenic characteristics identified. Results of Southern blot analysis indicate that the STa enterotoxin was preferentially encoded within an endemic plasmid of 4.9 Md. Other plasmids present in the E. coli isolates could be related to antibiotic resistance. With the exception of one strain, all E. coli isolates were resistant to more than one of the nine drugs tested; multiresistant E. coli were frequently isolated, including four strains which were resistant to seven antibiotics.     

Prevention of colibacillosis in neonatal swine with a 4-pilus E coli bacterin

Of 12 pregnant swine (vaccinates) given a 4-pilus enterotoxigenic E coli (ETEC) bacterin (K88, K99, 987P, F41), all developed comparable or significantly higher serum and colostral antibody levels than those of 8 pregnant swine (controls) given a 3-pilus ETEC bacterin (K88, K99, 987P). When piglets were challenged with an ETEC strain bearing the F41 antigen, those from vaccinates had significantly lower mortality, less scours, less severe clinical signs and better weight gain than those from controls.     

Serotyping of O and pilus antigens of Escherichia coli strains isolated from chickens with coli-septicemia.

One hundred and fifty one Escherichia coli strains were isolated from broiler chickens with coli-septicemia in Aichi (63 strains), Shizuoka (58 strains), and Kagoshima (30 strains) prefectures from 1980 to 1987, and their O and pilus antigens were serologically typed. One hundred and twenty five strains (82.8%) were typed into 23 O serogroups, and twenty six strains (17.2%) remained untypable. The predominant O serogroups were O2 (35 strains, 23.2%) and O78 (24 strains, 15.9%). Distribution of O serogroup was different, depending on prefectures where they are isolated. In total, 109 strains (72.2%) possessed Type 1 and/or Fmsha pili (Type 1; 41 strains, Fmsha; 22 strains, and Type 1 and Fmsha; 46 strains), and 42 strains (27.2%) were non-piliated. All the strains lacked K88, K99, 987P, F41, and Att25 pili. The ratios of piliated strains to non-piliated ones were almost the same among the three prefectures. Strains possessing Type 1 pili showed variety of O antigens, but most of the strains with Fmsha pili belonged to O2 serogroup.     

Protection of the nursing pig against experimentally induced enteric colibacillosis by vaccination of dam with fimbrial antigens of E coli (K88, K99 and 987P)

Pregnant gilts were vaccinated with two doses of alhydrogel adsorbed fimbrial antigens of Escherichia coli (K88ab, K88ac, K99 and 987P) supplemented with beta toxoid of Clostridium perfringens type C. Their piglets, and piglets of nonvaccinated gilts, were subsequently orogastrically challenged with one or other of the four fimbrial types of enteropathogenic E coli. Some of the vaccinated animals were reinjected with a single dose of the vaccine during second gestation and their piglets, and piglets of non-vaccinated sows, were challenged the same way as were litters of gilts. Blood serum and colostra were examined for antibodies to the four fimbrial antigens of E coli and for antitoxin to beta toxin of C perfringens type C. It was found that: (1) a highly significant reduction in mortality and morbidity was achieved in vaccinated litters against all four challenge strains of E coli; (2) excretion of K88ab and K88ac but not of K99 and 987P challenge strains was significantly reduced; (3) revaccination of sows by a single dose of the vaccine during second gestation conferred complete protection against mortality and highly significant protection against morbidity; (4) no correlation was noted between colostral or seroagglutinins to fimbrial antigens of E coli and mortality rates in litters challenged with homologous fimbrial types of E coli, but good correlation was found between colostral precipitins to K88 antigens and mortality rates in litters; (5) antitoxin value in 97 per cent of colostrum of vaccinated sows was 10 iu equivalent of C perfringens type C toxin or more per ml of colostrum.     

Use of enzyme-linked immunosorbent assay for detection of K88 pili in fecal specimens from swine

Enterotoxigenic Escherichia coli are important in neonatal pig diarrhea. To be pathogenic, E coli must possess the ability to produce toxin and adherence structures called pili. An enzyme-linked immunosorbent assay was used to detect K88 pili in feces of swine. Results were compared with standard isolation and identification techniques. Isolates were tested for labile toxin production by the Y-1 mouse adrenal gland test and presence of pili by the enzyme-linked immunosorbent assay test. The ability to detect K88 pili in feces correlated with the isolation of labile toxin-positive, K88-positive E coli.     

Detection of K88 and K99 fimbrial antigens on Escherichia coli by coagglutination

Coagglutination was used to detect K88 and K99 fimbrial antigens on Escherichia coli, and results were compared to an enzyme immuno assay (EIA). When pili suspensions were tested by both methods, 28 of 66 cultures were shown to have K88 and 11 of 31 cultures had K99 antigens. No pili suspensions were positive by coagglutination that were not positive by EIA. Testing of cell suspensions gave equivalent results to pili suspensions for K99 when tested by coagglutination. Two cell suspensions reacted with the K88 coagglutination which could not be confirmed by testing of pili suspensions, while a further 20 out of 43 cultures gave equivalent results with both cell and pili suspensions for K88 when tested by coagglutination.     

Prevalence of virulence genes in Escherichia coli strains recently isolated from young pigs with diarrhea in the US

Enterotoxigenic Escherichia coli (ETEC)-associated post-weaning diarrhea (PWD) is economically one of the most important diseases for the swine industry. Porcine ETEC strains typically express K88 or F18 fimbria and heat-labile (LT) and/or heat-stable (STa, STb) enterotoxins. However, recent studies indicate that EAST1 toxin, adhesin involved in diffuse adherence (AIDA-I) and porcine attaching and effacing-associated factor (paa) may also be expressed by ETEC strains associated with diarrhea. To better understand the virulence factors of E. coli strains that cause PWD, we applied PCR to screen for K88, F18, F41, 987P and K99 fimbrial genes; LT, STa, STb, Stx2e and EAST1 toxic genes; and AIDA-I, paa and EAE adhesin genes in E. coli strains recently isolated from young pigs with PWD in the US. Of 304 E. coli isolates from diarrheic pigs submitted for testing, 175 (57.6%) strains possessed fimbrial genes: K88 (64.6%), F18 (34.3%), F41 (0.57%), K99 (0.57%), 987P (0); toxin genes: LT (57.7%), STb (72.6%), STa (27.4%), STx2e (17.4%), EAST1 (35%); and adhesin genes: AIDA-I (26.9%), paa (60%), EAE (1.1%). All toxin genes except the EAST1 toxin gene, were almost exclusively associated with K88+ or F18+ isolates, and most of these isolates carried multiple toxin genes. The non-fimbrial adhesin paa was found present in over half of the K88+ isolates. A total of 129 (42%) isolates carried no fimbrial genes, including 66 (21.7%) isolates that did not have any of the above virulence genes. These results suggest a broad array of virulence genes associated with PWD in pigs.     

Attachment of E. coli-bearing K88 antigen to equine brush-border membranes

Equine small intestinal brush-border membranes, from 40 adult horses were tested in vitro for the presence of receptors for the Escherichia coli adhesive antigens K88ab, K88ac and K99. Only K88-positive strains of E. coli adhered strongly to horse brush-border membranes. In contrast, a K88-negative mutant strain J2, 2 K99-positive strains and 3 E. coli strains isolated from foals failed to adhere to horse brush-border membranes. Purified K88ac pili when reacted with equine brush-border membranes inhibited to a great extent the adhesion of K88-positive E. coli. Similarly, K88-positive E. coli previously reacted with K88 antibody, did not attach to equine brush-border membranes. Oral inoculation of 4 newborn foals with strains of K88-positive enterotoxigenic E. coli, producing either heat-stable or heat-labile enterotoxin, caused diarrhoea in 1 animal.