Introduction

The Pasteurellaceae family consists of several genera, of which Pasteurella pneumotropica biotypes “Jawetz” and “Heyl”, Actinobacillus muris, and Haemophilus influenzaemurium are regularly identified during routine health monitoring of laboratory mice. Pasteurellaceae constitute the most prevalent bacterial pathogens in experimental facilities worldwide [1,2]. Although the pathogenicity of most Pasteurellaceae species is low, P. pneumotropica is associated with variable clinical manifestations such as infections of the eye, genital tract, and respiratory system [3]. Even subclinical P. pneumotropica infections in immunocompetent mice could represent an unwanted experimental variable and may influence results [4]. Recently, the Federation of European Laboratory Animal Science Associations revised their recommendations on Pasteurellaceae reporting. While the prior version recommended monitoring all Pasteurellaceae species, specific identification of P. pneumotropica is currently recommended [5]. However, depending upon the research focus, some facilities may include identification of other Pasteurellaceae species, although this is controversial. Identification of Pasteurellaceae during routine health monitoring should be reliable, sensitive, fast, and easy to perform. However, no currently available method fulfills these criteria. Pasteurellaceae are usually detected by swab samples from the nasopharynx, genital tract, or large intestine and are cultured on agar plates, with subsequent analysis of suspicious colonies by using biochemical test kits. However, commonly used kits are optimized for human samples, and are not only unreliable for rodent sample identification to the species level but also sometimes fail to identify the correct bacterial family [5–7]. Analysis of subcultured bacterial colonies by matrix-assisted laser desorption/ionization-time of flight mass spectrometry offers good specificity; however, this method is expensive and lacks appropriate murine datasets. Additionally, serological tests are inappropriate for diagnosis of Pasteurellaceae infections, given that seroconversion of sentinel mice exposed to contaminated bedding or cohabitating with infected mice is unreliable [8]. There are also PCR-based options available for detection and differentiation of rodent-specific Pasteurellaceae: a genus-specific PCR assay to detect all Pasteurellaceae [9], species-specific assays to detect either of the P. pneumotropica biotypes “Jawetz” or “Heyl” [10,11], a multiplex PCR assay based on the 16S-23S rRNA internal-transcribed spacer region [12], and a real-time PCR assay that distinguishes between the P. pneumotropica biotypes “Heyl” and “Jawetz” [13]. Unfortunately, only a limited number of species can be identified by multiplex PCR assays, and the specific real-time PCR assay for P. pneumotropica “Jawetz” has a low discrimination potential against A. muris. PCR followed by fragment sequencing provides an accurate and reliable identification method. However, it is time consuming and expensive. The combination of real-time PCR assays with high-resolution melting curve analysis (HRM) [14] offers a relatively new molecular technique for genotyping, mutation scanning, sequence matching [15], and identification of bacterial species [16]. HRM analysis of the 16S rRNA gene, as well as other highly conserved genes, successfully identified microbial isolates, including Mycoplasma [17], Listeria [18–19], Brucella [20], Lactobacillus [21], and Staphylococcus [22] to the species or strain level. Recently, the first multiplex HRM real-time PCR assay was developed for simultaneous detection of Salmonella, Listeria monocytogenes, and Staphylococcus aureus [23]. The aim of this study was to develop a reliable, simple, and rapid alternative method to identify P. pneumotropica and other common Pasteurellaceae species by using a real-time PCR assay with HRM analysis.

Materials and Methods

Bacterial strains and isolates

Six Pasteurellaceae reference strains were used for assay development and as positive controls in each experiment (Table 1). Four were obtained from different culture collections (German Collection of Microorganisms and Cell Cultures: NCTC 8141 and NCTC 12432; Collection of the Pasteur Institute: CNP 160; Public Health England: NCTC 11146). The remaining two isolates included strains isolated and characterized within our facility (HMGU isolates). To evaluate the reference strains against a library of unknown isolates, 25 Pasteurellaceae isolates were used. These isolates were obtained from an external diagnostic laboratory and had been isolated from nasopharyngeal, tracheal, genital, or intestinal swab specimens from mice during routine health monitoring procedures.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Overview of reference species and assignment of clinical isolates.

https://doi.org/10.1371/journal.pone.0142560.t001

Results

HRM of reference strains

We assessed the suitability of the assay to distinguish between different Pasteurellaceae species and strains by testing the reference strains obtained from different culture collections or isolated from mice within our facility. Each reference strain exhibited a distinct HRM profile (Fig 1). The amplicons subjected to high-resolution melt generated melting peaks (T_m) ranging from 76.78°C to 79.10°C. The melting profiles of all strains resulted in a single peak with a maximum T_m variation of 0.1°C for each of the quadruplicates. Standard deviations within individual experiments ranged from 0.01 to 0.06 for all reference strains. Since the mean T_m values for the six strains differed by at least 0.21°C, they were deemed suitable for discriminating between the reference strains. Reproducibility was confirmed by analyzing the reference strains in four experiments (Table 1). The GC content of the six amplicons ranged from 43% to 47%, with amplicons having the same GC content displaying different melting points based on sequence differences. The P. pneumotropica “Heyl” and “HMGU isolate” reference strains differed in only one base pair (C/G–T/A). This single-base modification induced a 0.33°C shift in melting point.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. High-resolution melting analysis of six Pasteurellaceae reference strains.

Representative difference curves derived from the normalized data using the Actinobacillus muris (HMGU isolate) reference as the baseline.

https://doi.org/10.1371/journal.pone.0142560.g001

Discussion

In this study, we described a new approach using a real-time PCR/HRM system for identification and discrimination of Pasteurellaceae species, one of the most prevalent bacterial pathogens found in laboratory mice. The six reference species used for assay development exhibited sufficiently distinguishable melting profiles based on their different T_m values. Evaluation of the newly developed method using 25 unknown Pasteurellaceae isolates revealed accurate species identification of all specimens, confirmed by partial 16S rRNA sequencing. While sequencing is expensive and time consuming, the present method allows PCR and sequence analysis in a single step and can be completed in a few hours, thereby considerably increasing testing speed. Another advantage of this closed-tube method is reduced sample handling, resulting in elimination of PCR product contamination. Here, we used a short amplicon, because single-nucleotide modifications indicate unique melting points, even when specimens have the same GC content. Larger amplicons containing regions with different GC content may allow species differentiation based on changes in the melting curve profile. However, the effectiveness of detecting single-base substitutions decreases with increasing DNA-sequence length. In this study, sequence analysis revealed two opposite and adjacent single-nucleotide modifications in both H. influenzaemurium isolates as compared to the reference strain sequence. Species-level identification by HRM was reliable; however, this case revealed a major limitation of HRM. An unknown specimen not included in the reference library may possess the same T_m value as a non-associated reference strain. In this case, the melting point would allow assignment of the wrong species. This situation could only be avoided by rRNA sequencing, although accurate primer design might minimize the risk. Unknown strains with unprecedented melting profiles could be identified by sequencing and then be incorporated as additional reference controls. Here, only the application of pure isolated DNA from a single isolate yielded precise results. Based on these results, the combination of real-time PCR and HRM represents a fast and efficient application for the identification of Pasteurellaceae isolates grown from swab samples collected during health monitoring procedures. This simple and low-cost method has great potential to become a useful tool for research and diagnostic laboratories and can be easily adapted to other pathogenic bacteria.

Supporting Information