Development of a novel high resolution melting assay for identification and differentiation of all known 19 serovars of Actinobacillus pleuropneumoniae

Abstract Actinobacillus pleuropneumoniae is the etiological agent of porcine pleuropneumonia, a respiratory infectious disease responsible for global economic losses in the pig industry. From a monitoring perspective as well as due to the different courses of disease associated with the various serovars, it is essential to distinguish them in different herds or countries. In this study, we developed a novel high resolution melting (HRM) assay based on reference strains for each of the 19 known serovars and additional 15 clinical A. pleuropneumoniae isolates. The novel HRM comprises the species‐specific APP‐HRM1 and two serovar‐specific HRM assays (APP‐HRM2 and APP‐HRM3). APP‐HRM1 allowed polymerase chain reaction (PCR) amplification of apxIV resulting in an A. pleuropneumoniae specific melting curve, while nadV specific primers differentiated biovar 2 from biovar 1 isolates. Using APP‐HRM2 and APP‐HRM3, 13 A. pleuropneumoniae serovars can be determined by inspecting the assigned melting temperature. In contrast, serovar 3 and 14, serovar 9 and 11, and serovar 5 and 15 have partly overlapping melting temperatures and thus represent a challenge to accurately distinguish them. Consequently, to unambiguously ensure the correct assignment of the serovar, it is recommended to perform the serotyping HRM assay using a positive control for each serovar. This rapid and user‐friendly assay showed high sensitivity with 1.25 fg–125 pg of input DNA and a specificity of 100% to identify A. pleuropneumoniae. Characteristic melting patterns of amplicons might allow detecting new serovars. The novel HRM assay has the potential to be implemented in diagnostic laboratories for better surveillance of this pathogen.


| INTRODUCTION
Actinobacillus pleuropneumoniae is the etiological agent of porcine pleuropneumonia, a respiratory infectious disease responsible for global economic losses due to high mortality rates and high treatment costs (Gottschalk & Broes, 2019). A. pleuropneumoniae isolates can be divided into biovar I and biovar II, requiring exogenous nicotinamide adenine dinucleotide (NAD) for growth in the case of biovar I, whereas biovar II strains comprise nadV responsible for NADindependent growth (Pohl et al., 1983). Various serovars associated with different courses of disease are described, which can be differentiated by the expression of several capsular antigens (Sassu et al., 2018). Currently, there are 19 recognized serovars of A.
pleuropneumoniae based on the composition of the capsular polysaccharide (CPS) Stringer et al., 2021).
The pore-forming exotoxins ApxI, ApxII, and ApxIII are important virulence factors of A. pleuropneumoniae. All virulent strains express one or two of these toxins. ApxIV toxin is essential for full virulence of A. pleuropneumoniae. It is expressed by all isolates of this species making apxIV a useful species-specific marker (Chiers et al., 2010;Frey, 1995;Schaller et al., 1999).
However, using conventional PCR assays is more time-consuming and it remains a challenge to assign the correct amplicon size.
High resolution melting (HRM) is a rapid and low-cost PCR-based method characterizing PCR amplicons according to their dissociation behavior. Once a PCR reaction has been completed, a stepwise increase of temperature results in dissociation of the double-stranded DNA into single strands leading to a decrease in fluorescence intensity. The dissociation of the double-stranded DNA is dependent on the sequence of the amplicon, GC content, and length, therefore, contributing to a specific melting temperature for each amplicon (Vossen et al., 2009).
Serovar classification of A. pleuropneumoniae isolates helps trace certain serovars that cause severe diseases on a farm allowing epidemiological surveillance and is useful to provide information for vaccine development (Gottschalk, 2012). In this study, we propose a novel HRM assay to simultaneously identify A. pleuropneumoniae and its biovar based on apxIV and nadV on one hand and differentiate all known 19 serovars using CPS cluster as a target region on the other hand.

| Bacterial strains and clinical isolates
For the development of the novel HRM assay, the following A.

| Identification of clinical isolates
Serovars of clinical isolates were first identified using multiplex PCR of published protocols Stringer et al., 2021). Briefly, as proposed, two PCR reactions were performed using Qiagen HotStart Taq DNA Polymerase (Qiagen) in a minimal total reaction volume of 10 µl including primers at a final concentration of 0.3 μM each and 1 µl of DNA template. PCR cycling was performed with initial activation of Taq Polymerase for 15 min, followed by 35 cycles at 30 s for 94°C for denaturation, 90 s at 60°C, for annealing and 150 s at 72°C for elongation followed by a final extension step of 10 min at 72°C. Size analysis of PCR products was performed on a capillary electrophoresis QIAxcel Advanced device (Qiagen) using a screening cartridge, QX 15 bp-3 kb alignment marker, and QX 100 bp-2.5 kb size marker (Qiagen) according to the manufacturer's instructions. The resulting electropherograms were inspected with the QIAxcel ScreenGel 1.2.0 software (Qiagen).

| HRM development and optimization
Primers were designed using CLC Main Workbench software 7.5.1 (Qiagen) with CPS sequences of A. pleuropneumoniae retrieved from the NCBI databank targeting the same CPS loci as described previously (Stringer et al., 2021). The specificity of primer sequences was confirmed by BLAST searches. Oligonucleotide primers were T A B L E 1 Actinobacillus pleuropneumoniae reference strains used for the development of the high resolution melting (HRM) assays Nineteen reference strains (Table 1)

| Specificity
To determine the specificity of the HRM assay, an exclusivity panel of 27 bacterial isolates comprising 18 different bacterial strains was tested applying the three assays APP-HRM1, APP-HRM2, and APP-HRM3.

| Analytical sensitivity
To determine the analytical sensitivities of the HRM assay, all re- To analyze the range of detection and linearity of all 19 A.

| Efficiency
To calculate efficiencies of the HRM assays for each primer pair, C t values measured in triplicates were plotted against GE in form of standard curves using a 10-fold dilution series (5 × 10 6 GE, 5 × 10 5 GE, 5 × 10 4 GE, 5 × 10 3 GE, 500 GE, 50 GE, and 5 GE) of genomic DNA of each serovar reference strain. The PCR efficiency (E) was calculated from the slope (S) of the dilution curve in the linear range using the following equation:

| Specificity
The tested exclusivity panel of 27 pathogenic bacteria resulted in negative results for all tested non-A. pleuropneumoniae strains when performing APP-HRM1, APP-HRM2, and APP-HRM3. Furthermore, the reference strains of all 19 A. pleuropneumoniae serovars did not cross-react with other serovars (Figure 2). Hence, the novel HRM assay had a specificity of 100%.  (Figure 4) highlighting a low LOD of 5 GE with F I G U R E 1 High resolution melting (HRM) for identification of Actinobacillus pleuropneumoniae and biovar 2. APP-HRM1 assay allows targeting the species-specific gene apxIV for identification of A. pleuropneumoniae and nadV for biovar 2 detection, respectively.

| Analytical sensitivity
A. pleuropneumoniae strains N273 (serovar reference strain 13), 3906 (serovar reference strain 14), SS4935 (serovar 2), and SS4936 (serovar 2) (represented in blue) all contain apxIV and nadV, whereas all remaining A. pleuropneumoniae strains tested in the study (represented in red) are biovar 1 and therefore only harbor apxIV linearity of the standard curve across a large range of DNA quantities between 5,000,000 GE and 5 GE.
For APP-HRM2 and APP-HRM3 the linear range of standard curves was more variable in contrast to APP-HRM1 due to the increased complexity of the master mixes containing up to 10 cpsspecific primer-pairs. Obtained LODs within the relevant confidence level of 95% were 5 GE for A. pleuropneumoniae serovar 5b; 50 GE for A. pleuropneumoniae serovar 4, 5a, 10, and 16; 500 GE for A.
pleuropneumoniae serovar 1, 2, 3, 7, 8, 9, 11, 14, 18, and 19; and 5000 GE for A. pleuropneumoniae serovar 6, 12, 13, 15, and 17, respectively. Standard curves had correlation coefficients of R 2 > 0.96 (Table 6).   space unit (Stygar et al., 2016). Due to the challenges of antibiotic usage and the generation of antibiotic resistance, the most promising approach to preventing A. pleuropneumoniae infection lies in vaccination (Cao et al., 2020;Michael et al., 2015). Current market-leading vaccines are based on inactivated Apx toxins and outer membrane components of A. pleuropneumoniae or F I G U R E 3 Identification of Actinobacillus pleuropneumoniae serovar 9 and serovar 11 using high resolution melting (HRM) assay APP-HRM3. Differentiation of A. pleuropneumoniae serovar 9 and serovar 11 is based on a single-nucleotide polymorphism in cps9/11F. A 10-fold dilution series of A. pleuropneumoniae reference strains serovar 9 (light green) and serovar 11 (dark green) was tested in triplicates. Illustration of (a) PCR amplification curves, (b) melting curves of the HRM step, (c) normalized plot, and (d) difference plot normalized with genomic DNA from A. pleuropneumoniae serovar 11 (12.5 ng) visualizing two groups corresponding to A. pleuropneumoniae serovars 9 and 11. PCR, polymerase chain reaction The primer pairs targeting serovar 9 and 11 can differentiate the serovars since the amplified PCR product encompasses a single-nucleotide polymorphism (SNP) in cps 9/11F leading to slightly different T m of the corresponding PCR amplicons. Representation as a difference plot allows visualizing the slight difference in the melting curve. Due to missing clinical isolates representing these serovars, no validation using field isolates could be fulfilled. It is recommended in the future to test more isolates with serovar 9 and 11 of different origins for validation purposes and to prove the ability to robustly discriminate these closely related serovars.

| Clinical isolates
When performing APP-HRM1, PCR amplification of apxIV resulted in a melting curve with a T m of 71.8 ± 0.2 unambiguously identifying A. pleuropneumoniae, while the melting curve obtained by nadV specific primers yielded T m of 79.8 ± 0.2 determining biovar 2 isolates. Using APP-HRM2 and APP-HRM3, 13 A. pleuropneumoniae serovars (serovars 1, 4, 7, 8, 13, 16, 17, 18, and 19) could be determined explicitly inspecting the assigned T m listed in Tables 3-5. In contrast, serovar 3 and 14, serovar 9 and 11, and serovar 5 and 15 cannot explicitly be distinguished uniquely from the T m . Serovar 3 (T m = 75.7 ± 0.2) and serovar 14 (T m = 75.5 ± 0.1), serovar 9 (73.3 ± 0.1) and serovar 11 (73.4 ± 0.1), serovar 5 (T m = 74.5 ± 0.1) and serovar 15 (T m = 74.8 ± 0.2), respectively, harbor partly overlapping T m . Additionally, serovars 2 (T m = 77.6 ± 0.1) and 10 (T m = 77.3 ± 0.1) and similarly, serovars 6 (T m = 77.6 ± 0.1) and 12 (T m = 77.2 ± 0.1) have close T m values. Since this newly proposed HRM assay is based on the high resolution melting of PCR amplicons, which directly depends on its sequences, unique melting temperatures are expected for each serovar. It is a challenge to visualize the 19 different APP serovars in an HRM setting of only two reaction mixes. To unambiguously ensure the correct F I G U R E 4 Identification of Actinobacillus pleuropneumoniae targeting apxIV (APP-HRM1) illustrating high sensitivity. APP-HRM1 was performed with an A. pleuropneumoniae tenfold dilution series of genomic DNA (here as an example A. pleuropneumoniae serovar 15) using primers targeting apxIV. DNA quantities of the tenfold dilution series were 5,000,000 genome equivalents (GE) (red), 500,000 GE (orange), 50,000 GE (yellow), 5000 GE (green), 500 GE (light blue), 50 GE (blue), and 5 GE (violet). Representation of the 10-fold dilution series as (a) qPCR amplification plot; (b) melting curves of the HRM step illustrating a limit of detection of 5 GE; (c) standard curve indicating linearity across a large range of DNA quantities between 5,000,000 GE and 5 GE with a high correlation coefficient (R 2 > 0.99). HRM, high resolution melting; qPCR, quantitative polymerase chain reaction T A B L E 6 Efficiency and limit of detection (LOD) of APP-HRM1, APP-HRM2, and APP-HRM3 targeting apxIV, nadV, and serovar-specific cps loci Note: The LOD of APP-HRM1 was between 5 and 50 genome equivalents (GE) corresponding to 12.5-125 fg of genomic DNA with PCR efficiencies of 93%-105%. LODs of APP-HRM2 and APP-HRM3 were between 5 and 5000 GE corresponding to 12.5 pg-12.5 fg genomic DNA with PCR efficiencies of 90%-108%.
Abbreviations: APP-HRM, Actinobacillus pleuropneumoniae-high resolution melting; LOD, limit of detection; PCR, polymerase chain reaction. Serovar prevalence differs from country to country. In England and Wales serovar 8 is most prevalent , whereas serovar 7 plays an important role in Spain (Maldonado et al., 2009).
Interestingly, several studies from central Europe reported clear dominance of serovar 2, as described in recent studies from Germany and Hungary Schuwerk et al., 2021) and partially outdated ones from Belgium, Denmark, and the Netherlands (Dom et al., 1994;Jessing et al., 2003). Furthermore, in countries on other continents, such as Canada and Australia serovar 5, 7, and 15, respectively, were most frequently detected (Gottschalk & Lacouture, 2015;Turni et al., 2014

CONFLICTS OF INTEREST
The authors declare no conflicts of interest.

ETHICS STATEMENT
None required.

DATA AVAILABILITY STATEMENT
All data are provided in this article except for the supplemental data, which are available in the Zenodo repository at https://doi.org/10. 5281/zenodo.6045373 (Table S1: Inter-and intra-assay variability of APP-HRM1 and Table S2: Inter-and intra-assay variability of APP-HRM2 and APP-HRM3).