A curated public database for multilocus sequence typing (MLST) and analysis of Haemophilus parasuis based on an optimized typing scheme

Haemophilus parasuis causes Glässer's disease and pneumonia in swine. Serotyping is often used to classify isolates but requires reagents that are costly to produce and not standardized or widely available. Sequence-based methods, such as multilocus sequence typing (MLST), offer many advantages over serotyping. An MLST scheme was previously proposed for H. parasuis but genome sequence data only recently available reveals the primers recommended, based on sequences of related bacteria, are not optimal. Here we report modifications to enhance the original method, including primer redesign to eliminate mismatches with H. parasuis sequences and to avoid regions of high sequence heterogeneity, standardization of primer T(m)s and identification of universal PCR conditions that result in robust and reproducible amplification of all targets. The modified typing method was applied to a collection of 127 isolates from North and South America, Europe and Asia. An alignment of the concatenated sequences obtained from seven target housekeeping genes identified 278 variable nucleotide sites that define 116 unique sequence types. A comparison of the original and modified methods using a subset of 86 isolates indicates little difference in overall locus diversity, discriminatory power or in the clustering of strains within Neighbor-Joining trees. Data from the optimized MLST were used to populate a newly created and publicly available H. parasuis database. An accompanying database designed to capture provenance and epidemiological information for each isolate was also created. The modified MLST scheme is highly discriminatory but more robust, reproducible and user-friendly than the original. The MLST database provides a novel resource for investigation of H. parasuis outbreaks and for tracking strain evolution.