felis or A. genospecies 2. After 3 days (corresponding to an increase

of OD600 nm from 0.02 to 0.6), the culture was diluted to an OD600 nm of 0.02. This step was repeated at least seven times. Bacteria were grown on agar without antibiotics and five single colonies were selected as templates for colony PCR. The primer pair MCS-2 FP01 and MCS-2 RP01 (Table 1) led to the production of a single polymerization product HSP inhibitor at approximately 800 bp in the presence of the plasmid. Cover slips were coated with poly-l-lysine and left with bacteria in PBS for 30 min at an ambient temperature. Nonattached bacteria were removed by rinsing three times with PBS and samples were fixed with 3% formaldehyde in PBS for 40 min. GFP-bacteria were visualized at 488 nm excitation and 522 nm emission. Monoclonal antibody CSD11 or

rabbit serum were used compound screening assay as primary antibodies in immunofluorescence, together with Alexa488-coupled goat-anti-rabbit or goat-anti-mouse antibodies as secondary antibodies. Slides were mounted with mowiol, examined and photographed using an Axiophot Epifluorescence Microscope (Zeiss, Oberkochen, Germany). Following a published protocol (Riess et al., 2003) for transposome-directed mutagenesis in the closely related B. henselae using commercially available transposome technology, between 450 and 1900 mutants were obtained per microgram of transposome DNA, making this approach extremely costly. Published efficiencies obtained with this transposition system varied from 1200 clones μg−1 DNA with Xylella fastidiosa (Koide et al., 2004) to 107 clones μg−1 DNA with enteric bacteria (Hoffman et al., 2000). As the electrotransformation efficiency of

A. felis using plasmid DNA was within the expected range, one explanation for the low efficiency was the digestion Doxacurium chloride of the introduced DNA fragments by an Afipia DNA restriction system. Recently, a purified phage protein called ‘ocr’ (Walkinshaw et al., 2002) became available. This phage protein is a strong inhibitor for type I endonucleases (Murray, 2000). Adding purified inhibitor to the transformation mixture increased the efficiency from ∼2000 kanamycin-resistant clones per microgram of transposome DNA to >3 × 104 (Fig. 1). Although it is not known whether Afipia spp. have type I restriction enzyme systems, the strong increase in transposon mutant yields using the inhibitor suggests that the transposon sequence contained a restriction site that is recognized by a type I restriction endonuclease of Afipia. Electroporation in the absence of a transposome yielded no colonies, as expected. To test whether all kanamycin-resistant Afipia clones contained a transposon, we performed PCR reactions with the primer pair Tnp FP01/Tnp RP01 internal of the transposon yielding 1109-bp DNA fragments in positive cases. Eighty-five of 86 tested clones contained a transposon.