High-Resolution Melting-Curve Analysis of obg Gene to Differentiate the Temperature-Sensitive Mycoplasma synoviae Vaccine Strain MS-H from Non-Temperature-Sensitive Strains

Temperature-sensitive (ts +) vaccine strain MS-H is the only live attenuated M. synoviae vaccine commercially available for use in poultry. With increasing use of this vaccine to control M. synoviae infections, differentiation of MS-H from field M. synoviae strains and from rarely occurring non-temperature-sensitive (ts –) MS-H revertants has become important, especially in countries where local strains are indistinguishable from MS-H by sequence analysis of variable lipoprotein haemagglutinin (vlhA) gene. Single nucleotide polymorphisms (SNPs) in the obg of MS-H have been found to associate with ts phenotype. In this study, four PCRs followed by high-resolution melting (HRM)-curve analysis of the regions encompassing these SNPs were developed and evaluated for their potential to differentiate MS-H from 36 M. synoviae strains/isolates. The nested-obg PCR-HRM differentiated ts + MS-H vaccine not only from field M. synoviae strains/isolates but also from ts – MS-H revertants. The mean genotype confidence percentages, 96.9±3.4 and 8.8±11.2 for ts + and ts – strains, respectively, demonstrated high differentiating power of the nested-obg PCR-HRM. Using a combination of nested-obg and obg-F3R3 PCR-HRM, 97% of the isolates/strains were typed according to their ts phenotype with all MS-H isolates typed as MS-H. A set of respiratory swabs from MS-H vaccinated specific pathogen free chickens and M. synoviae infected commercial chicken flocks were tested using obg PCR-HRM system and results were consistent with those of vlhA genotyping. The PCR-HRM system developed in this study, proved to be a rapid and reliable tool using pure M. synoviae cultures as well as direct clinical specimens.


Introduction
Mycoplasma synoviae causes airsacculitis and infectious synovitis in chickens and turkeys [1]. It causes significant economic losses to the poultry industry due to carcass condemnation, culling of lame birds and deterioration in eggshell quality [2,3]. The temperaturesensitive (ts + ) strain MS-H (Vaxsafe MSH, Bioproperties Pty. Ltd. Australia) is the only live attenuated vaccine available and is used in several countries to control M. synoviae infections in poultry flocks.
Differentiation of MS-H from field strains is an important step to establish whether a flock is free from wild-type M. synoviae. It is also important to establish whether the vaccine strain has colonised the respiratory mucosa so as to produce an efficient immune response to protect against wild-type disease. A number of PCR-based techniques have been reported for typing of M. synoviae strains, targeting the vlhA gene [4][5][6][7], 16S rRNA genes [8] or the 16S to 23S rRNA intergenic spacer region [9,10]. Only a small number of these studies included the MS-H vaccine in their experiments. Jeffery et al. [11] described a combination of PCR and high-resolution melting (HRM) curve analysis of the vlhA gene products to discriminate a large number of M. synoviae strains, although their system did not differentiate MS-H from several Australian field strains as they shared the same vlhA gene sequence. The vlhA-based typing system however should be useful in other countries as MS-H like strains are believed to be rare, if not absent, outside Australia. Pulsed-field gel electrophoresis (PFGE) using BlnI and BamHI digestions coupled with vlhA gene sequencing was useful in differentiating the MS-H from Japanese M. synoviae strains/isolates [12]. Also a PCR based cycling probe technology (CPT) developed by Ogino et al. [13], targeting an ARG substitution at 365 th nucleotide from the 59 conserved region of vlhA gene, has been claimed useful for MS-H differentiation from Japanese M. synoviae strains/isolates. However techniques reported in both of these reports are time consuming and may be difficult to perform on a routine basis in diagnostic laboratories. More importantly, none of the techniques reported Pure culture [15] MS-H (+) V,G1 Australia, vaccine strain derived from 86079/7NS Pure culture [15] MS-H P5 (+) V,G1 MS-H reisolate, Australia (with 5 consecutive in-vivo passages) Pure culture This study MS-H 4 (2) V,G2 MS-H reisolate, Australia Pure culture [14] MS-H 5 (2) V,G2 MS-H reisolate, Australia Pure culture [14] MS-H 3 (2) V,G2 MS-H reisolate, Australia Pure culture [14] 94036/10-5a (2) V,G2 MS-H reisolate, Australia Pure culture [14] 93198/5-10a (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93205/1-2a (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93205/2-9a (2) V,G2 MS-H reisolate, Australia Pure culture [14] 93198/6-5b (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93198/4-19a (+) V,G1 MS-H reisolate, Australia Pure culture [14] 94036/2-2a (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93198/3-13b (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93198/3-15a (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93205/2-13a (2) V,G2 MS-H reisolate, Australia Pure culture [14] 93205/9-3a (+) V,G1 MS-H reisolate, Australia Pure culture [14] 93205/8-9c ( above have the capacity to distinguish between MS-H and its nontemperature sensitive isolates rarely isolated from vaccinated flocks [14]. Microtitration followed by incubation at two different temperatures has been used to determine the temperature-sensitive (ts) phenotypes of M. synoviae strains/isolates [15]. We have recently developed a technique using a combination of differential growth at two different temperatures with a quantitative real-time PCR (vlhA Q-PCR) to determine ts phenotype of M. synoviae strains [16] however this technique still requires culture of the organism and therefore access to live cloned organism.
We have recently compared partial genome sequences of MS-H, its parent strain 86079/7NS and two ts -MS-H reisolates (MS-H 4 and MS-H 5 ) and found an SNP (GRA) at nucleotide position 367 in MS-H obg, an essential gene encoding highly conserved GTP binding protein Obg found in organisms ranging from human to bacteria. Obg is involved in essential cellular processes such as signal transduction, protein synthesis, ribosome biogenesis, DNA replication initiation, chromosomal segregation and progression through cell cycle [17]. A nucleotide change (GRA) at position 367 in MS-H obg, causing an alteration of glycine to arginine at position 123 in Obg fold, was predicted to play a role in temperature sensitivity phenotype of MS-H [18]. Analysis of complete obg nucleotide sequences from further 19 MS-H reisolates revealed another SNP (CRT) at position 629, causing amino acid change from alanine to valine at position 210 in GTP binding domain, in 4 MS-H reisolates [18]. These SNPs were used in this study to develop a rapid and reliable test, using HRM-curve analysis, to differentiate MS-H from ts -MS-H reisolates and/or M. synoviae field strains.

Ethics statement
Clinical swab samples were taken from palatine cleft, trachea or sinus of specific pathogen free (SPF) chickens, vaccinated with M. synoviae vaccine strain MS-H, after euthanasia using intravenous injection of phenobarbitone as per approval of the Melbourne University Animal Ethics Committee (Approval number 0911472.1). Swabs from field commercial chicken flocks were submitted as diagnostic specimens.

Mycoplasma strains and growth media
All mycoplasma strains used in this study are listed in Table 1. A total of 36 M. synoviae strains/isolates and 28 clinical swab samples were used in this study. The collection of MS-H reisolates examined in this study (Table 1) is a unique collection prepared in our laboratory through extensive monitoring of the MS-H vaccinated flocks. For 23 out of 36 M. synoviae strains/isolates, the ts phenotype was determined in a previous study [16]. For MS-H P5, WVU-1853, 94036/5-5a and 94036/6-3a, the ts phenotype was determined in this study using microtitration as described before [16]. Other 9 field strains, characterised as different from MS-H based on the analysis of PCR products from vlhA gene single-copy conserved region by single strand conformation polymorphism and HRM genotyping [11] (Table 1), were presumed ts -. Especially since natural ts + M. synoviae strains have not been reported so far, examination of temperature sensitivity of the field strains was not considered in this study. The ts + phenotype is a property of the MS-H vaccine produced by chemical mutagenesis [15,[19][20][21]. Where necessary, M. synoviae strains/ isolates were grown in mycoplasma broth (MB) containing 10% swine serum and 0.01% (w/v) of nicotinamide adenine dinucleotide (NAD) [22]. M. gallisepticum vaccine strain ts-11 was used as control to test the specificity of obg PCR primers.

DNA extraction
Swabs taken from MS-H vaccinated SPF chickens and nonvaccinated commercial chicken flocks were subjected to DNA extraction. Cell pellet of 500 ml MB culture of M. synoviae strains/ isolates, harvested by centrifugation at14,0006g for 1 min, or respiratory swabs were placed in 500 ml Qiagen RLT lysis buffer containing 1% of 2-b-mercaptoethanol and incubated at 4uC overnight. After a brief vortex, the swab (where applicable) was removed and 15 ml of Qiaex II matrix (Qiagen, Chadstone, Victoria, Australia) and 300 ml of 70% ethanol added to the lysis buffer. The suspension was mixed and loaded onto a multispin MSK-11 column (Axygen, Union City, California, USA) and placed in a 1.5 ml microfuge tube and centrifuged for 30 sec at 10,0006g with the flow-through discarded. Columns were washed once with 600 ml of RW1 buffer (Qiagen) and twice with 500 ml of RPE buffer (Qiagen) followed by centrifugation for 30 sec at 10,0006g after each wash. The spin column was dried by centrifugation for 90 sec at 14,0006g. Finally, 50 ml of nuclease free water was added to the columns and DNA was eluted after incubation at room temperature for 5 min and centrifugation at 10,0006g for 60 sec. Similar amount of DNA (,50 ng/ml) was used in all experiments although this was less controllable for clinical specimens submitted as swabs. Extracted DNA was used immediately in PCR or stored at 220uC for future use.

Oligonucleotide primers
The nucleotide primers used in this study, and their sequences, are listed in Table 2 while their location are shown on Figure 1A. All primers were designed using AmplifX version 1.5.4 and PerlPrimer version 1.1.20 [23]. The primers obg-F1 and obg-R1 were designed to flank obg SNP 367. Primers obg-F3 and obg-R3 were designed to flank the obg SNP 629. Primers obg-F and obg-R were designed for partial sequencing of the obg from M. synoviae strains/isolates. Specificity of primers was evaluated using BLAST search against non-redundant nucleotide databases.
obg PCRs Three regions of the obg, encompassing both or one of the SNPs detected in MS-H (GRA and CRT at positions 367 and 629, respectively), were targeted by PCR for HRM-curve analysis ( Figure 1A). The obg-F1R3 PCR spanned over both SNPs while the obg-F1R1 and obg-F3R3 PCRs spanned over SNP GRA or CRT, respectively. PCR reactions were carried out in iCycler thermal cycler (Bio-Rad, Gladesville, New South Wales, Australia). A 25 ml PCR reaction mixture contained 1 ml each of 25 mM forward and reverse oligonucleotides (0.1 ml each of 25 mM oligonucleotides for obg-F3R3 PCR), 2 ml of 25 mM MgCl 2 , 4 ml of 1.25 mM dNTP mixture, 1 U of GoTaqH DNA polymerase (Promega, Alexandria, New South Wales, Australia), 5 ml of 56GoTaqH flexi green buffer (Promega), 2 ml of 100 mM SYTO 9 green fluorescent nucleic acid stain (Invitrogen, Mount Waverley, Victoria, Australia), 1 ml of M. synoviae genomic DNA (,50 ng/ml) and 10.8 ml of nuclease free water. PCR reaction conditions for F1R1 PCR included an initial denaturation at 95uC for 2 min, and then 35 cycles of 95uC for 10 sec, 50uC for 20 sec and 72uC for 25 sec. PCR reaction conditions for obg-F1R3 PCR included an initial denaturation at 95uC for 2 min, and then 35 cycles of 95uC for 30 sec, 49uC for 30 sec and 72uC for 30 sec. PCR reaction conditions for obg-F3R3 PCR included an initial denaturation at 95uC for 2 min, and then 35 cycles of 95uC for 30 sec, 58uC for 30 sec and 72uC for 15 sec. In each set of reaction, nuclease free water was used as negative control. All specimens were tested in triplicates.
To amplify a 841-bp region of the obg for sequencing purposes, the obg-FR PCR was conducted using oligonucleotide primers obg-F and obg-R.  ation at 95uC for 2 min then 45 cycles of 95uC for 10 sec, 48uC for 10 sec and 72uC for 60 sec.
All PCR products were analysed by electrophoresis through 1% agarose gels stained with GelRed TM (Biotium, Hayward, California, USA) and visualised by UV transillumination.

High-resolution melting-curve analysis
High-resolution melting-curve analysis was conducted in a Rotor-Gene 6000 thermal cycler (Corbett Life Science, Mortlake, New South Wales, Australia) and signal detected using an excitation wavelength at 470 nm and detection at 510 nm. Melting-curves were generated by increasing the temperature from 60 to 90uC for obg-F1R1, obg-F3R3 and obg-F1R3 PCR products and recording the fluorescence. To optimise melting conditions for maximum differentiation of sequence differences, PCR products were subjected to different ramp speeds of 0.05, 0.1, 0.2, 0.3 and 0.5uC per sec. The HRM-curve analysis was performed using the software Rotor-Gene 1.7.27 and HRM algorithm provided. Conventional melt-curves were generated automatically. To generate normalised HRM-curves, following normalisation regions were applied: 72.5 to 73.0 and 77.5 to 78.0 for obg-F1R1; 70.9 to 71.9 and 79.2 to 80.2 for obg-F3R3 and 74.5 to 76.0 and 80.5 to 82.0 for obg-F1R3. The MS-H profile was set as 'genotype' and the average HRM genotype confidence percentages (C%) (value attributed to each strain being compared to the genotype, with a value of 100 indicating an exact match) for replicates were automatically calculated by Rotor-Gene 1.7.27. The C% value attributed to all other strains/isolates indicated similarity of the given strain/isolate to the ts + MSH. The mean C% of specimen replicates and standard deviations were calculated using Microsoft TM Office Excel 2003.

Nested-obg PCR-HRM
Oligonucleotide primers obg-F1 and obg-Ri2 were used to amplify a 60-bp internal region of obg (harbouring the SNP 367) from products generated in obg-F1R1 PCR ( Figure 1A). PCR was performed in 25 ml reaction volumes containing 5 ml of 56Go-TaqH flexi green buffer (Promega), 0.1 ml each of 25 mM oligonucleotide primers, 2 ml of 25 mM MgCl 2 , 2 ml of 1.25 mM dNTP mixture (Promega), 0.2 ml of GoTaqH DNA polymerase (Promega), 2 ml of 100 mM SYTO 9 green fluorescent nucleic acid stain (Invitrogen), 2 ml of 0.016 diluted obg-F1R1 PCR product as template and 10.8 ml of nuclease free water. PCR conditions consisted of denaturation at 95uC for 2 min followed by 35 cycles of 95uC for 10 sec, 52uC for 20 sec and 72uC for 10 sec. All reactions were carried out in triplicate. In each experiment, water instead of template was used as negative control and MS-H and parent strain 86079/7NS genomic DNA were used as ts + and tscontrols, respectively. Following PCR, HRM-curve analysis was carried out in Rotor-Gene 6000 thermal cycler (Corbett Life Science Pty Ltd) as described above. Melting curves were generated by increasing the temperature from 66 to 78uC at ramp speeds of 0.1, 0.2, 0.3 and 0.5uC per sec. Normalisation regions of 66.5 to 67.0 and 76.5 to 77.0 and genotype confidence threshold of 85% were applied to characterise unknown M. synoviae strains/isolates using MS-H as genotype/reference strain.
Nucleotide sequencing and sequence analysis PCR products generated in obg-FR PCR (841-bp) were separated through 1% agarose gels, bands of expected size were excised, purified using WizardH SV Gel and PCR clean-Up System (Promega) and cloned into pGEMH-T Easy vector (Promega) using instructions provided by the manufacturer. The resultant constructs were propagated in a-select competent cells, silver efficiency (Bioline, Alexandria, New South Wales, Australia) and extracted using PureYield TM Plasmid Miniprep (Promega). All purified PCR products or plasmid extracts were subjected to automated sequencing (BigDye Terminator v3.1; Applied Biosystems, Foster City, California, USA) in both directions using primers obg-F and obg-R, or M13 forward and reverse sequencing primers for purified PCR products or cloned PCR products, respectively. Nucleotide sequences were edited using SeqMan TM II and EditSeq programs in DNASTAR. Multiple sequences were aligned using computer program ClustalW2. Nucleotide sequence of complete obg of 25 strains including MS-H and its related strains, belonging to four different genotypes, has been described in our previous study [18]. Nucleotide sequence of partial obg of additional 10 M. synoviae strains, including 94041/12a, 4GPH3, F10-2AS, K1938, K870, K1858, YA, K1968, K1723 and WVU-1853, has been submitted to GenBank under accession numbers KF875990 to KF875999.

PCR amplification of selected regions of the obg from different M. synoviae strains/isolates
In order to evaluate the capacity of obg PCRs for HRM analysis to differentiate MS-H from M. synoviae strains, five sets of oligonucleotide primers, as detailed in Table 2 and Figure 1A, were used to amplify 5 regions of 841, 335, 101, 99 and 60-bp of obg from four M. synoviae strains/isolates. All strains/isolates generated PCR products of the expected size in all PCRs as confirmed by agarose gel electrophoresis ( Figure 1B). No PCR product was detected from M. gallisepticum strain ts-11 DNA and no template negative control ( Figure 1B) indicating specificity of the obg PCRs.
The obg-F1R3 PCR-HRM curve analysis could not reliably differentiate MS-H from M. synoviae strains/isolates tested The 335-bp obg-F1R3 PCR products from 16 M. synoviae strains/isolates including MS-H and its related isolates, and a number of field strains from Australia and the USA were subjected to HRM-curve analysis ( Figure S1). Conventional melt-curve analysis of the PCR products using a ramp of 0.3uC/sec showed that all strains generated a single peak at 79.760.1uC which were also visually very similar in pattern making it difficult to differentiate MS-H from other strains ( Figure S1 and Table S1). Visual examination of the normalised HRM-curves also showed very minor differences between curve profiles of MS-H and other M. synoviae strains/isolates ( Figure S1). When genotyping was applied to the normalised HRM-curves using MS-H as reference genotype, the C% 6 SD for the strains/isolates 93198/1-24b, 94036/5-5a, 4GPH3, K870, WVU-1853 and YA were 78.8610.0, 83.463.8, 71.861.6, 87.968.8, 89.460.5 and 83.2612.2, respectively. All these strains could be auto-called as 'variation' from MS-H when a genotype confidence threshold of 90% was applied. For other strains/isolates, the C% was above 90% (93.763.6) and the normalised melt curves were mostly similar to that of MS-H on visual examination. Therefore the obg-F1R3 HRM-curve analysis was not considered as a reliable tool and was not pursued any further in this study for differentiation of MS-H from other strains/isolates.

The obg-F1R1 PCR-HRM curve analysis differentiated MS-H from all M. synoviae field strains/isolates but not from WVU-1853
The 101-bp obg-F1R1 PCR products, spanning over SNP GRA at position 367, from various M. synoviae strains/isolates were subjected to HRM-curve analysis. Only a small number of strains/isolates were used in this assay to provide a preliminary evaluation of the assay. Visual examination of conventional melt curves at different ramps revealed that a ramp of 0.3uC/sec generated the most distinct curves and therefore used in the further HRM analysis. The conventional melting-curve analysis showed a single peak for all strains examined. The melting peaks for MS-H vaccine, its ts + reisolates, and the US strain WVU-1853 occurred at 75.860.0uC while those for all other strains including 86079/7NS and ts -MS-H reisolates occurred at 76.360.1 ( Figure  S1). Normalised HRM-curve analysis distinctly separated strains into two groups, one for the known ts + and the other for the tsstrains with the exception of the rarely occurring ts -MS-H reisolates with mutation at position 629 ( Figure S1). When genotyping with a C% threshold of 90 was applied, two distinct genotypes were auto-called: one included MS-H and ts + MS-H reisolates (mean C%, 97.262.6) and the other included tsstrains (mean C%, 6.164.7). The MS strain WVU-1853 and the rarely occurring ts -MS-H reisolates had normalised HRM-curves identical to that of MS-H ( Figure S1). HRM data from different experiments is shown in Table S2.

Alignment of partial obg nucleotide sequences revealed further SNPs
Alignment of partial nucleotide sequence of obg from MS-H, MS-H reisolates (both ts + and ts -) and field M. synoviae strains/ isolates revealed further nucleotide variations in obg (Figure 2), especially in the region targeted in obg-F1R1 PCR. M. synoviae strains F10-2AS, K1723, YA and WVU-1853 had CRT variation at position 402 while 94041/12a had CRA variation at position 434. Therefore, a further oligonucleotide primer (obg-Ri2) was designed to allow targeting of the region spanning over the SNP GRA at position 367 and avoiding other polymorphic sites found in obg. Nucleotide sequence alignment of obg regions targeted in nested-obg and obg-F3R3 PCR-HRM, for all M. synoviae strains/ isolates used in this study except field isolate 94036/8-3a, is shown in Figure S2 and S3, respectively. For reasons unknown to the authors, several attempts at sequencing the obg-FR PCR product for 94036/8-3a were failed. Similarly, attempts at sequencing the vlhA region of 94036/8-3a were unsuccessful in our previous study [16].  revealed that all known ts + MS-H reisolates generated HRMcurves similar to those for MS-H while the ts -M. synoviae strains/ isolates and ts -MS-H reisolates (except 93198/1-24b, 94036/9-2a and 93198/6-1a) had HRM-curves similar to those for 86079/ 7NS. After applying genotyping to the normalised curves using a C% threshold of 84, two distinct genotypes were auto-called: the MS-H type with a mean C% of 96.963.4, and variants with a mean C% of 8.8611.2. MS-H reisolates (with a tsphenotype) that previously could not be differentiated from MS-H, due to identical vlhA region, were distinguishable from MS-H in the nested-obg PCR-HRM (Table 3).
Nested-obg HRM melting points for MS-H, 86079/7NS and other strains grouped with either of them exhibited minor variation in melting temperature on different days using different DNA extractions as templates but the melting point differences between MS-H and 86079/7NS remained $0.7uC (Table S3). Normalised HRM-curves, in all instances, correctly genotyped all strains/isolates either with MS-H or 86079/7NS. Furthermore, all tested US strains (F10-2AS, K1723, K1858, K1938, K1968, K870, YA and WVU-1853) were autocalled as variant from MS-H genotype and produced melting-curves (73.160.0uC) and C% (5.161.8) identical to 86079/7NS, and therefore, characterised as ts - (Figure 3A and B and Table 3). HRM data for all (36) M. synoviae strains/isolates, used in this study, is shown in Table S3.

Nested-obg PCR-HRM curve analysis successfully applied for direct examination of clinical specimens
The nested-obg PCR-HRM was first optimised using DNA extracted from pure cultures of M. synoviae strains/isolates as described above, and then extended to clinical swab specimens taken from sinus, palatine cleft or trachea of SPF and field chickens inoculated intra-ocularly with MS-H. Swabs from palatine cleft and trachea of non-vaccinated commercial chicken flocks were used as negative control. All swabs from MS-H   vaccinated SPF chickens produced single melting peak at 72.060.0uC while swab samples from non-vaccinated commercial field chicken flocks generated peak at 72.860.0uC. Melting peaks for MS-H and 86079/7NS as ts + and tscontrols were 72.060.0uC and 72.860.0uC, respectively, thus visual examination of the melting-curves could clearly differentiate the two different melting profiles ( Figure 3C). Application of genotyping on normalised curves (using MS-H as reference), distinctively classified the specimens either as MS-H (mean C% of 99.360.6) or variation (4.161.3) ( Figure 3D). Therefore there was an approximate gap of 95% in C% for these two groups. HRM data for all swab samples tested in this study is shown in Table S3.
Thus, coupling of nested-obg with obg-F3R3 PCR-HRM had 100% accuracy in differentiation of MS-H from all field strains and ts -MS-H reisolates. Also a high accuracy (97.2%) was achieved in predicting the ts phenotype of M. synoviae strains/ isolates ( Table 5). Irrespective of the (unknown) prevalence of strains with identical nucleotide at position 629 to that of MS-H (e.g., 86079/7NS), the presence of isolates with SNP at position 629 reflects that obg-F3R3 PCR is more useful when combined with the nested obg-PCR.

Discussion
SNPs are the most common type of genetic variation and have been used for species/strain identification of various bacterial pathogens [24][25][26][27][28][29]. A large number of methods have been utilised for rapid identification of SNPs. These include hybridisation-based methods (molecular beacons, SNP microarrays), enzyme-based methods (restriction fragment length polymorphism, flap endonuclease, primer extension, 59-nuclease, oligonucleotide ligase assay) and post-amplification methods based on physical properties of the DNA (single strand conformation polymorphism, temperature gradient gel electrophoresis, high resolution melting analysis) [30]. Among these, HRM is thought to be rapid and at the same time most economical where a small number of specimens are to be genotyped [31,32].
In this study, a set of SNPs detected by comparative genomic sequence analysis of M. synoviae strains of MS-H lineage were The potential of nested-obg PCR-HRM was initially evaluated using pure M. synoviae cultures available in our laboratory. In order to evaluate the potential of the nested-obg PCR-HRM directly on clinical specimens, two sets of clinical swabs, one from MS-H vaccinated SPF birds and the other from commercial chicken flocks suspected (by serological monitoring) of M. synoviae infection were tested. The swabs from MS-H vaccinated SPF birds generated HRM-curves identical to that of MS-H culture while swab samples from infected field birds produced clearly different pattern, identical to that generated by 86079/7NS, indicating infection by a field strain at the time of sampling. This demonstrated the potential of the nested-obg PCR-HRM to rapidly determine whether a flock is infected with a field M. synoviae strain or it harbours MS-H vaccine strain. Evaluation of the full potential of this assay for direct examination of clinical specimens should include determination of its sensitivity and specificity although it should be noted that the primary use of this assay in our laboratory has been confined to examination of pure (cloned) M. synoviae cultures.
The nested-obg HRM also discriminated ts + MS-H from tsfield strains and ts -MS-H reisolates. Out of total 36 M. synoviae strains/isolates, 33 were typed according to their ts phenotype. The ts phenotype of three MS-H reisolates (93198/1-24b, 93198/ 6-1a and 94036/9-2a) was not determined in accordance with their temperature sensitivity phenotype (genotyped as 'MS-H' by nested-obg PCR-HRM but exhibited tsphenotype). Complete obg sequence of these three strains and one MS-H reisolate 94036/2-1a (ts + ) revealed a secondary mutation (CRT) at position 629. Therefore, an alternative PCR-HRM (obg-F3R3 PCR-HRM), encompassing SNP at position 629 was developed to discriminate these rare variants from all other M. synoviae strains/isolates. It is recommended that where an unknown M. synoviae strain/isolate genotyped as MS-H by the nested-obg HRM, the obg-F3R3 HRM should be used to determine its ts phenotype and identity. Thus, a combination of nested-obg and obg-F3R3 HRM-curve analysis not only differentiates MS-H from field M. synoviae strains and from ts -MS-H reisolates, but also exhibited high accuracy (97.2%, 35/36) in predicting the ts phenotype of any unknown M. synoviae strain/isolate. The exception was the rare isolate 94036/2-1a (ts + ) with obg mutations at position 367 and 629. The influence of this second mutation on ts phenotype of M. synoviae has been discussed in our previous report [18].
Temperature-sensitive bacterial mutants have been produced only in laboratories, mostly by N-methyl-N-nitro-nitrosoguanidine (NTG) [15,[19][20][21]. Such ts + mutants are expected to sustain more than one ts mutation contributing to overall ts phenotype and accounting for the genetic stability of temperature-sensitive mutants [20]. Therefore obg SNPs based assays developed in this study may not be ideal for ts phenotyping of other organisms although due to their consistency for the MS-H and its reisolates, were found highly useful for M. synoviae genotyping purposes.
Wild strain mutation causing ts + (MS-H-like phenotype) has never been reported to the best of our knowledge. However reversion from a ts + to tsphenotype is expectable under favourable selective pressure [15,33]. The back mutation rate of the MS-H vaccine strain from a ts + phenotype to tsphenotype was found in the order of 10 -4 . No study on the virulence and transmissibility of ts + MS-H reisolates have been conducted although a previous study in our laboratory demonstrated that ts -MS-H reisolates did not have the characteristics, including virulence potential, of the vaccine parent strain [33] and that factors other than ts phenotype may be involved in loss of virulence of MS-H. Nevertheless the recovery of tsstrain/isolate from healthy vaccinated flocks prompted the current study to establish the true identity of the isolate.
The combination of nested-obg and obg-F3R3 PCR-HRM is relatively rapid and can be completed in one day after DNA extraction. High discriminating power of this genotyping system, with an added advantage of predicting the ts phenotype, makes it an ideal assay that can be routinely used in veterinary diagnostic laboratories involved in M. synoviae genotyping especially in countries where MS-H is routinely used in commercial poultry. Figure S1 High-resolution melting-curve analysis of M. synoviae strains/isolates using obg-F1R3 and obg-F1R1 PCR products.