Supporting Materials and Methods

Microarray Hybridizations.

The Massachusetts Institute of Technology Helicobacter hepaticus strain collection comprises strains isolated in the US (11 strains), The Netherlands (one strain), and Germany (one strain). All of these were used for comparison with the sequenced strain by microarray hybridizations (1). The identification of all strains as H. hepaticus was confirmed by 16S rDNA sequence analysis and phenotypical characterization. The MWG H. hepaticus array (MWG Biotech, Ebersberg, Germany) consists of 50-mer oligonucleotides permitting the detection of 1,863 of the 1,875 ORFs. The design strategy for MWG arrays has been described (2). Fluorescent labelling of DNAs and competitive hybridizations were essentially performed as described by Salama et al. (3), with the following modifications: genomic DNA was prepared from plate-grown bacteria by a genomic Tip-100 kit (Qiagen, Valencia, CA). Labeling of genomic DNA with aminoallyl-UTP was done with the BioPrime DNA Labeling System (Invitrogen). One microgram of DNA was diluted in a final volume of 24 m l H2O, mixed with 20 m l of random primer solution and heated at 99°C for 5 min. Afterwards, 5 m l of dNTP mix (0.5 mM dGTP, dATP, and dCTP, 0.3 mM dTTP, and 0.2 mM aminoallyl-dUTP) and 1 m l of Klenow was added and the reaction was incubated for 1 h. The reactions were stopped by adding 5 m l of stop mix and purified by using the Qiagen PCR Purification kit with the following modifications: wash steps were done with phosphate wash buffer (5 mM KPO4, pH 8.0, 80% EtOH) and elution took place with 2´ 30 m l of phosphate elution buffer (4 mM KPO4, pH 8,5). Labeling with Cy-dyes and the following purification were done as in the protocol by Salama et al., except that the Cy3 and Cy5 reactions were combined after purification and 1 m l of yeast-tRNA was added. After drying the probe was resuspended in 25 m l of hybridization buffer (50% formamide, 6´ SSC, 0.5% SDS, 50 mM Na-phosphate, pH 8.0, 5´ Denhardt’s solution). The probe was heated for 2 min at 99°C and applied to the microarray. Microarray scanning and data processing were performed as described previously (2). The program gack (4), which uses the signal-ratio distribution rather than a fixed cutoff, was used to label genes as "present" or "absent". The total or partial absence of HHGI1 from seven H. hepaticus isolates was confirmed by PCR analyses using primers in ORFs flanking the island (empty site PCR), as well as primers targeting representative ORFs within the island. The primer sequences are available upon request. The precise location of the deletion was determined by sequencing of the empty site PCR product. The absence of representative ORFs from the other strains was similarly verified.

1. Saunders, K. E., McGovern, K. J. & Fox, J. G. (1997) J. Clin Microbiol. 35, 2859-2863.

2. Josenhans, C., Niehus, E., Amersbach, S., Hörster, A., Betz, C., Drescher, B., Hughes, K. T. & Suerbaum, S. (2002) Mol. Microbiol. 43, 307-322.

3. Salama, N., Guillemin, K., McDaniel, T. K., Sherlock, G., Tompkins, L. & Falkow, S. (2000) Proc. Natl. Acad. Sci. USA 97, 14668-14673.

4. Kim, C. C., Joyce, E. A., Chan, K. & Falkow, S. (2002) Genome Biol. 3, RESEARCH0065.