Downloaded from www.microbiologyresearch.org by IP: 54.146.60.16 On: Sun, 30 Jul 2017 12:46:07 J. Med. Microbiol. - Vol. 41 (1994), 63-68 0 1994 The Pathological Society of Great Britain and Ireland Genomic DNA fingerprinting of clinical Haemophilus infhenzae isolates by polymerase chain reaction amplification : comparison with major outer-membrane protein and restriction fragment length polymorphism analysis A. VAN BELKUM", BIRGITTA DUIMT, ANNETTE REGELINK?, LIEKE MOLLERJ-5, WIM QUINT*$ and L. VAN ALPHEN? * Diagnostisch Centrum SSDZ, Department of Molecular Biology, PO Box 50 10, 2600 GA Delft; -/- University of Amsterdam, Academic Medical Center, Department of Medical Microbiology, Meibergdreef 15, 1 105 AZ Amsterdam; 1 University Hospital Dijkzigt, Department of Virology, Dr. Molewaterplein 40, 3015 GD Rotterdam and Q Laboratory for Public Health Groningen and Department of Medical Microbiology, University Hospital Groningen, Oostersingel 59, 9713 EZ Groningen, The Netherlands Summary. Non-capsulate strains of Haemophilus injluenzae were genotyped by analysis of variable DNA segments obtained by amplification of genomic DNA with the polymerase chain reaction (PCR fingerprinting). Discrete fragments of 100-2000 bp were obtained. The reproducibility of the procedure was assessed by comparing : (i) the fingerprints of 16 colonies of a single H. injluenzae strain; (ii) isolates obtained from individual sputum samples (a total of 57 H. injluenzae isolates from three cystic fibrosis patients); and (iii) 17 isolates collected during an outbreak of H. injluenzae infection in a local pulmonary rehabilitation centre. The discriminatory power of the method was demonstrated by showing that the PCR fingerprints of eight unrelated H. injluenzae strains from sputum samples of patients with chronic obstructive pulmonary disease (COPD) and 32 strains from cystic fibrosis patients were all different. These 40 isolates also differed with respect to their restriction fragment length polymorphisms (RFLP) and major outer-membrane protein (MOMP) composition. Twelve MOMP antigenic strain variants from sputum samples of five COPD patients had identical PCR fingerprints and RFLPs. It was concluded that PCR fingerprinting is a reliable and reproducible method for genotyping non-capsulate strains of H . injluenzae. The discrimi- natory power of PCR fingerprinting was similar to that of RFLP analysis, but the results of PCR fingerprinting were easier to interpret. Introduction Haemophilus injluenzae is found exclusively in man. Non-capsulate strains cause acute and persistent infec- tions of the mucous membranes and colonise the nasopharynx of a high percentage of healthy in- dividuals.' Throat cultures of children can remain positive for a particular H. injluenzae strain for several months2* In patients with chronic obstructive pul- monary disease (COPD) and cystic fibrosis (CF), H. injluenzae is isolated frequently from sputum samples.4* However, the availability of high- resolution bacterial typing assays is a pre-requisite for the study of H. injluenzae epidemiology. Non-capsulate H. injluenzae isolates are charac- Correspondence to Dr A. van Belkum, Academic Hospital Dijkzigt, Department of Clinical Microbiology, Dr Molewaterplein 40, 301 5 GD Rotterdam, The Netherlands. Received 20 Dec. 1993; accepted 28 Jan. 1994. terised commonly by major outer-membrane protein (MOMP) analysis with SDS-PAGE. This method is highly discriminatory because of strong inter-strain However, MOMP patterns have been shown to change during persistent H. influenzae infections in patients with COPD.l0.'l Antigenic di- versity of lipopolysaccharides (LPS), which may also be used to characterise non-typable H. injluenzae strains,12 is difficult to interpret because LPS under- goes phase variation at high frequency.13 Of the genotypic methods, multilocus enzyme electro- phoresis,'* rRNA gene restriction pattern analysis'' and DNA restriction fragmevt length polymorphism (RFLP) analysis have been applied successfully to differentiate isolates.l03 11, l6 However, since multilocus enzyme electrophoresis is very laborious, rRNA gene restriction pattern analysis has a low discriminatory power, and RFLP analysis results in over-complex banding patterns, there is a clear need for a novel discriminatory typing met hod.