Single knock-outs of ivy, mliC and pliG as well as an ivy/mliC double knock-out were successfully constructed in APEC CH2, and plasmid-based complementation of the mutants with the corresponding genes was accomplished. First we determined the serum resistance of the mutants as a rapid and simple indicator of virulence, and found that mliC, but not ivy or pliG, was required for serum resistance of APEC CH2. Although bacterial sensitivity to serum is mainly due to the action of the complement system, there is also a contribution of other antimicrobial components such as lysozyme. The action of the membrane attack complex of the complement system destabilizes the outer membrane and may render it permeable to lysozyme. Conversely, degradation of the peptidoglycan layer may facilitate pore formation in the cytoplasmic membrane by the membrane attack complex, resulting in cell leakage an death. ubsequently, infection experiments with 1-day old chickens subcutaneously injected with different doses of bacteria confirmed the attenuated virulence of the mliC mutant. In addition, virulence was fully restored by complementation with the mliC gene. As BEZ235 915019-65-7 anticipated from the serum resistance test, pliG nor ivy had any significant effect on virulence. Since PliG is the only known inhibitor of g-type lysozyme in APEC, and its knock-out reduced g-type lysozyme inhibitory XAV939 284028-89-3 activity of APEC CH2 to background levels, it can be concluded that PliG is not required for virulence of this pathogen, at least not in the subcutaneous infection model used in this work. Of course, a role of this inhibitor in other commensal or pathogenic bacteria – host interactions can not be excluded on the basis of these observations. For the c-type lysozyme inhibitors, the situation is more complex. Based on the observations with the single knock-out strains, the outer membrane-bound inhibitor MliC appears to play a role in virulence, but not the periplasmic inhibitor Ivy. Since MliC is an outer membrane protein, there could be some concern that knock-out of MliC could have destabilized the outer membrane, thus rendering the bacteria more sensitive to a variety of antibacterial effectors from its host. This appears not to be the case, because the mliC mutant retained its resistance to detergents when plated on LB containing 2.0% SDS or 2.0% Triton X-100, whereas mutants with outer membrane defects typically display a high serum and detergent sensitivity. Therefore, we can have confidence that the attenuated virulence of the mliC mutant is genuinely linked to its reduced production of c-type lysozyme inhibitor rather than to an indirect effect. One point that needs further clarification is which 
inhibitor is responsible for the attenuated virulence, since the mliC mutant unexpectedly showed a considerably reduced level of periplasmic lysozyme inhibitor activity. An additional complication, in line with the observations in the serum resistance test, is that introduction of an ivy knock-out into the mliC mutant restored the attenuated virulence of the latter to almost wild-type level again, indicating that there is some type of interference between these two mutations. Comparison of the periplasmic lysozyme inhibitory activities confirms that this is indeed the case, because the level in the double mutant is higher than that in the mliC mutant. For comparison, an ivy mliC mutant of E. coli MG1655 was previously shown to have no residual periplasmic lysozyme inhibitory activity, but an explanation for this strain-dependent behaviour is currently lacking.