Characterization of repetitive sequences controlling phase variation of Haemophilus influenzae lipopolysaccharide.

Phase variation of lipopolysaccharide epitopes of an Haemophilus influenzae serotype b strain (strain RM.7004) occurs through a mechanism which depends on multiple tandem repeats of the DNA sequence 5'-CAAT-3' situated within the chromosomal locus lic1. We report here that the same tetranucleotide repeats are also found in two other genomic loci (lic2 and lic3) of RM.7004. Similar to lic1, there are multiple tandem repeats of 5'-CAAT-3' present at the 5' ends of long open reading frames in lic2 and lic3. Variation in the number of repeats of CAAT, by shifting the upstream initiation codons in or out of phase with the remainder of the open reading frame, could switch on or off the translation of downstream genes. Similar to previously reported findings for lic1, site-directed mutations in the open reading frame downstream (3') from the repeats of CAAT in lic2 abolished phase variation and identified DNA sequences required for the expression of additional oligosaccharide epitopes. When we used an oligonucleotide comprising five repeats of CAAT or DNA sequences specific for lic1, lic2, and lic3 as probes, a survey of other encapsulated H. influenzae strains (serotypes a through f) and nontypable H. influenzae strains (including biotype aegyptius) showed that the chromosome of H. influenzae can have from two to five regions which contain multiple tandem repeats of CAAT in addition to other sequences which hybridize to lic1 and lic2.