Escherichia coli are major bacterial pathogens causing bovine mastitis, a disease of great economic impact on dairy production worldwide. This work aimed to study the virulence determinants of mammary pathogenic E. coli (MPEC). By whole-genome sequencing analysis of 40 MPEC and 22 environmental (“dairy-farm” E. coli [DFEC]) strains, we found that only the fec locus (fecIRABCDE) for ferric dicitrate uptake was present in the core genome of MPEC and that it was absent in DFEC genomes (P < 0.05). Expression of the FecA receptor in the outer membrane was shown to be citrate dependent by mass spectrometry. FecA was overexpressed when bacteria were grown in milk. Transcription of the fecA gene and of the inner membrane transport component fecB gene was upregulated in bacteria recovered from experimental intramammary infection. The presence of the fec system was shown to affect the ability of E. coli to grow in milk. While the rate of growth in milk of fec-positive (fec+) DFEC was similar to that of MPEC, it was significantly lower in DFEC lacking fec. Furthermore, deletion of fec reduced the rate of growth in milk of MPEC strain P4, whereas fec-transformed non-mammary gland-pathogenic DFEC strain K71 gained the phenotype of the level of growth in milk observed in MPEC. The role of fec in E. coli intramammary pathogenicity was investigated in vivo in cows, with results showing that an MPEC P4 mutant lacking fec lost its ability to induce mastitis, whereas the fec+ DFEC K71 mutant was able to trigger intramammary inflammation. For the first time, a single molecular locus was shown to be crucial in MPEC pathogenicity.

Mastitis remains a major infection of dairy cows and an important issue for dairy farmers and the dairy industry, in particular infections due to Escherichia coli strains. So far, properties specific to E. coli causing mastitis remain ill defined. In an attempt to better understand the properties required for E. coli to trigger mastitis, we used a range of in vitro assays to phenotypically characterize four E. coli strains, including the prototypical E. coli mastitis strain P4, possessing different relative abilities to cause mastitis in a mouse model. Our results indicate that a certain level of serum resistance might be required for colonization of the mammary gland. Resistance to neutrophil killing is also likely to contribute to a slower clearance of bacteria and higher chances to colonize the udder. In addition, we show that the four different strains do induce a pro-inflammatory response by mammary epithelial cells but with different intensities. Interestingly, the prototypical mastitis strain P4 actually induces the less intense response while it is responsible for the most severe infections in vivo. Altogether, our results suggest that different strategies can be used by E. coli strains to colonize the mammary gland and cause mastitis.

Escherichia coli is one of the major pathogens causing mastitis in dairy cattle. Yet, the factors which mediate the ability for E. coli to develop in the bovine mammary gland remain poorly elucidated. In a mouse model, infections induced by the reference mastitis E. coli P4 showed a strong colonisation of the mammary gland, while this strain had a low stimulating power on cells of the PS bovine mammary epithelial cell line. In order to understand if such a reduced response contributes to the severity of infection, a library of random mutants of P4 strain was screened to identify mutants inducing stronger response of PS cells. Among hyper-stimulating mutants, six were altered in genes involved in biosynthesis of lipopolysaccharide (LPS) and had lost their O-polysaccharide region, suggesting that the presence of O-antigen impairs the response of PS cells to LPS. Using purified smooth (S) and rough (R) fractions of LPS, we showed that the R-LPS fraction induced a stronger response from PS cells than the smooth LPS fraction. Biological activity of the S-LPS fraction could be restored by the addition of recombinant bovine CD14 (rbCD14), indicating a crucial role of CD14 in the recognition of S-LPS by Mammary Epithelial Cells (MEC). When S-LPS and R-LPS were injected in udder quarters of healthy lactating cows, an inflammation developed in all infused quarters, but the S-LPS induced a more intense pro-inflammatory response, possibly in relation to sizeable concentrations of CD14 in milk. Altogether, our results demonstrate that the O-antigen modulates the pro-inflammatory response of MEC to LPS, that S-LPS and R-LPS trigger different responses of MEC and that these responses depend on the presence of CD14.

The outer membrane protein (Omp) A is a major constituent of the outer membrane of Escherichia coli. This protein has been used in several vaccine development studies, but seldom with a view to vaccinating against mastitis. The objective of this study was to investigate the immunogenicity of E. coli OmpA and its vaccine potential for cows. Both the humoral and cellular immune responses were investigated. The gene for OmpA of the mastitis-causing strain P4 was cloned and expressed, and the recombinant protein (rEcOmpA) purified. Cows were immunized twice with rEcOmpA with adjuvant one month apart by the systemic route. Before immunization, few antibodies to rEcOmpA were detected, and there was little production of IL-17A in a whole blood stimulation assay (WBA) with rEcOmpA. Antibodies to rEcOmpA were induced by immunization. These antibodies were not able to react with E. coli P4, but reacted with a rough P4 mutant prepared by inactivating the rfb locus. This suggests that the complete LPS O-chain precluded the accessibility of antibodies to their target at the outer membrane. The cellular immune response appeared to be biased towards a Th17-type, as more IL-17A than IFN-γ was produced in the OmpA-specific WBA. There was a good correlation between antibody titers and the production of IL-17A in the WBA. The intramammary instillation of rEcOmpA elicited a slight local inflammatory response which was not related to the WBA. Overall, the interest of OmpA as vaccine immunogen was not established, although other experimental conditions (dose, adjuvant, route) need to be investigated to conclude definitively. The study pointed to several important issues such as the accessibility of OmpA to antibodies and the weakness of Th1-type response induced by OmpA.

Escherichia coli is the leading cause of severe mastitis in dairy farms. As E. coli mastitis is refractory to the hygienic control measures adapted to contagious mastitis, efficient vaccines are in demand. Existing mastitis vaccines, based on the use of killed rough E. coli J5 as the antigen, aim at inducing phagocytosis by neutrophils. We assessed the binding of J5-induced antibodies to isogenic rough and smooth strains along with a panel of mastitis-associated E. coli. Analysis by enzyme-linked immunosorbent assay revealed that antibodies to OmpA or killed J5 bind readily to rough E. coli but poorly to smooth strains. Flow cytometry analysis indicated that immunization with J5 induced antibodies that cross-reacted with rough E. coli strains but with only a small subpopulation of smooth strains. We identified type 1 fimbriae as the target of most antibodies cross-reacting with the smooth strains. These results suggest that the O-polysaccharide of lipopolysaccharide shields the outer membrane antigens and that only fiber antigens protruding at the bacterial surface can elicit antibodies reacting with mastitis-associated E. coli. We evaluated J5-induced antibodies in an opsonophagocytic killing assay with bovine neutrophils. J5 immune serum was not more efficient than preimmune serum, showing that immunization did not improve on the already high efficiency of naturally acquired antibodies to E. coli. In conclusion, it is unlikely that the efficiency of J5 vaccines is related to the induction of opsonic antibodies. Consequently, other research directions, such as cell-mediated immunity, should be explored to improve E. coli mastitis vaccines. IMPORTANCE Despite intensive research, mastitis remains an important disease in dairy cattle with a significant impact on animal welfare, use of antibiotics, and, in the end, the economy of dairy farms. Although vaccines available so far have shown limited efficacy against coliform mastitis, vaccination is considered one of the measures that could limit the consequences of mastitis. One reason for the lack of efficiency of current vaccines likely stems from the current evaluation of vaccines that relies mostly on measuring antibody production against vaccine antigens. This report clearly shows that vaccine-induced antibodies fail to bind to most mastitis-associated E. coli strains because of the presence of an O-antigen and, thus, do not allow for improved phagocytosis of pathogens. As a consequence, this report calls for revised criteria for the evaluation of vaccines and suggests that cell-mediated immunity should be targeted by new vaccinal strategies. More generally, these results could be extended to other vaccine development strategies targeting coliform bacteria.