Novel DNA Markers for Identification of Actinobacillus pleuropneumoniae.

Actinobacillus pleuropneumoniae causes porcine pleuropneumonia, an important disease in the pig industry. Accurate and sensitive diagnostics such as DNA-based diagnostics are essential for preventing or responding to an outbreak. The specificity of DNA-based diagnostics depends on species-specific markers. Previously, an insertion element was found within an A. pleuropneumoniae-specific gene commonly used for A. pleuropneumoniae detection, prompting the need for additional species-specific markers. Herein, 12 marker candidates highly conserved (99 - 100% identity) among 34 A. pleuropneumoniae genomes (covering 13 serovars) were identified to be A. pleuropneumoniae-specific in silico, as these sequences are distinct from 30 genomes of 13 other Actinobacillus and problematic [Actinobacillus] species and more than 1700 genomes of other bacteria in the Pasteurellaceae family. Five marker candidates are within the apxIVA gene, a known A. pleuropneumoniae-specific gene, validating our in silico marker discovery method. Seven other A. pleuropneumoniae-specific marker candidates within the eamA, nusG, sppA, xerD, ybbN, ycfL, and ychJ genes were validated by polymerase chain reaction (PCR) to be specific to 129 isolates of A. pleuropneumoniae (covering all 19 serovars), but not to four closely related Actinobacillus species, four [Actinobacillus] species, or seven other bacterial species. This is the first study to identify A. pleuropneumoniae-specific markers through genome mining. Seven novel A. pleuropneumoniae-specific DNA markers were identified by a combination of in silico and molecular methods and can serve as additional or alternative targets for A. pleuropneumoniae diagnostics, potentially leading to better control of the disease. IMPORTANCE Species-specific markers are crucial for infectious disease diagnostics. Mutations within a marker sequence can lead to false-negative results, inappropriate treatment, and economic loss. The availability of several species-specific markers is therefore desirable. In this study, 12 DNA markers specific to A. pleuropneumoniae, a pig pathogen, were simultaneously identified. Five marker candidates are within a known A. pleuropneumoniae-specific gene. Seven novel markers can be used as additional targets in DNA-based diagnostics, which in turn can expedite disease diagnosis, assist farm management, and lead to better animal health and food security. The marker discovery strategy outlined herein requires less time, effort, and cost, and results in more markers compared with conventional methods. Identification of species-specific markers of other pathogens and corresponding infectious disease diagnostics are possible, conceivably improving health care and the economy.

INTRODUCTION

Porcine pleuropneumonia is an important disease with high economic impact for the swine industry (1, 2). Economic loss from the disease is attributed to pig mortality, reduction in daily weight gain, a longer rearing period, lower feed efficiency, as well as medication and veterinary expenses (1, 2). Porcine pleuropneumonia affects pigs of all ages. The disease can be acute with fibrino-hemorrhagic and necrotizing pneumonia, leading to sudden death (3, 4). Pigs that survive acute infection or recover after remedial treatment may become disease carriers (3, 4). It is therefore important to monitor pigs for pleuropneumonia to ensure that they remain free of the disease to promote animal health, food security, and the economy.

The causative agent of porcine pleuropneumonia is Actinobacillus pleuropneumoniae, a Gram-negative bacterial pathogen of the pig respiratory tract. This species currently consists of 19 serovars (5), which can be distinguished mainly by unique capsular polysaccharide (CPS) antigens, as lipopolysaccharide O-antigens (LPS O-Ags) can be shared by groups of serovars such as 1/9/11, 3/6/8/15 and 4/7 (6, 7). Despite some genomic differences among various serovars, core genes exist (8) and potentially contain species-specific DNA markers.

A. pleuropneumoniae diagnostics are important for surveillance, prevention, and control of porcine pleuropneumonia. Effective diagnostics can guide decisions on antibiotic treatment, quarantine, and vaccine usage. Diagnosis based on clinical signs can be unreliable, as symptoms may be common to various respiratory diseases. The ability to correctly identify and distinguish the species of interest from closely related species is important for guiding an appropriate response to a disease outbreak. DNA-based detection methods such as polymerase chain reaction (PCR) can be highly specific, allowing discrimination of different species when the targeted DNA sequences are sufficiently unique. Amplification of A. pleuropneumoniae-specific DNA in pig-derived samples (e.g., lung tissues, nasal swabs, tonsils, and oral fluids) is therefore exploited for disease diagnosis (9–12).

Many DNA markers and PCR assays for A. pleuropneumoniae detection have been reported (3, 5, 9, 13–15). Some assays, however, have limitations regarding their specificity, as they are unable to distinguish A. pleuropneumoniae from closely related Actinobacillus species (3, 5, 13–16). Assays based on the apxIVA gene, encoding a repeats-in-toxin (RTX) family protein, are A. pleuropneumoniae-specific (9), making this gene an excellent target for A. pleuropneumoniae detection. However, mutations within species-specific markers, especially at or within primer binding sites, can lead to diagnostic evasion (5, 17–19). An example of serodiagnostic escape in A. pleuropneumoniae is the AP76 strain which contains the ISApl1 insertion element in the apxIVA gene. The insertion element disrupts the gene, ablates ApxIV expression, and prevents ApxIV-based serological detection (Tegetmeyer et al., 2008). Depending on the primers used, such insertions can affect the results of apxIVA-based PCR assays, possibly leading to misinterpretation (5, 17). The availability of multiple species-specific markers is therefore desirable to ensure accurate detection and prevent diagnostic evasion.

Previously, A. pleuropneumoniae-specific markers were discovered empirically by cross-species hybridization or PCR in which DNA fragments that can serve as species-specific markers were identified (13, 15, 16, 20). Now, with growing numbers of genome sequences of various pathogens available in public databases, these genome assemblies can be utilized for identification of new species-specific DNA markers for diagnostic purpose. Using genome sequence data to identify species-specific markers is superior to empirical testing of DNA fragments, since the content of whole genome can be screened comprehensively in silico, covering more putative markers and potentially yielding more species-specific markers. In this study, whole-genome sequences of A. pleuropneumoniae were mined for novel A. pleuropneumoniae-specific markers. The new markers identified can serve as alternative or additional markers in A. pleuropneumoniae-specific diagnostics.

RESULTS

In silico identification of novel A. pleuropneumoniae-specific DNA marker candidates.

In order to identify new A. pleuropneumoniae-specific DNA markers, 11 complete A. pleuropneumoniae genome assemblies (Table 1) with sizes ranging between 2.24 – 2.41 Mb covering 7 serovars (serovars 1–5, 7, and 8) were analyzed. A. pleuropneumoniae-conserved sequences of 100 – 400 nucleotides sharing 100% identity among the 11 genomes were identified. Using MegaBLAST searches against the nucleotide collection (nr/nt) database and the WGS database of Pasteurellaceae, which include 34 A. pleuropneumoniae genomes covering 13 serovars, 30 genomes of 13 other Actinobacillus and [Actinobacillus] species, 116 genomes of G. parasuis, and 291 genomes of P. multocida (Table 2), 12 A. pleuropneumoniae-conserved sequences were shown to be specific to A. pleuropneumoniae in silico (Table 3). These 12 sequences are called “A. pleuropneumoniae-specific marker candidates.” Each of the marker candidates are highly conserved among A. pleuropneumoniae genomes (Table 3). Five A. pleuropneumoniae-specific marker candidates are within the apxIVA gene, a known A. pleuropneumoniae-specific marker (9, 20), validating our in silico marker identification method as effective. In addition to the five apxIVA sequences, seven sequences also fit the criteria of being A. pleuropneumoniae-specific marker candidates. These seven sequences are within the eamA, nusG, sppA, xerD, ybbN, ycfL, and ychJ genes (Table 3). Nucleotide sequences of these markers are shown in supplemental material. The presence and specificity of these marker candidates were further validated by PCR.

TABLE 1

Accession numbers of A. pleuropneumoniae complete genome assemblies used for identification of A. pleuropneumoniae-conserved sequences

No.	Strain	Serovar	Accession no.	Genome size (Mb)	Reference
1	ATCC 27088T	1	CP030753.1	2.32	(47)
2	ATCC 27088T	1	CP029003.1	2.32	(8)
3	KL16	1	CP022715.1	2.37	(48)
4	CCUG 47657	2	LR134515.1	2.33
5	JL03	3	CP000687.1	2.24	(49)
6	ATCC 33378	4	LS483358.1	2.34
7	L20	5b	CP000569.1	2.27	(50)
8	App6	5	CP026009.1	2.41
9	AP76	7	CP001091.1	2.35
10	MIDG2331	8	LN908249.1	2.34	(51)
11	405	8	CP078508.1	2.32

ATCC, american type culture collection; CCUG, culture collection university of gothenburg; T indicates type strain of the species.

TABLE 2

Number of genome assemblies of selected species from the Pasteurellaceae family or of other pig pathogens in the NCBI databases available for in silico comparison

Species	No. of total genome assemblies(in the nr/nt and WGS databases)	No. of complete genome assemblies(in the nr/nt database)
Actinobacillus capsulatus	1	0
[Actinobacillus] delphinicola	1	1
Actinobacillus equuli	3	3
[Actinobacillus] indolicus	3	1
Actinobacillus lignieresii	3	1
[Actinobacillus] minor	2	0
Actinobacillus pleuropneumoniae	34	11
[Actinobacillus] porcinus	2	0
[Actinobacillus] porcitonsillarum	1	1
[Actinobacillus] seminis	2	0
[Actinobacillus] succinogenes	1	1
Actinobacillus suis	7	2
Actinobacillus ureae	3	0
Actinobacillus vicugnae	1	0
Aggregatibacter actinomycetemcomitans	97	12
Bibersteinia trehalosi	7	4
Escherichia coli	24529	1782
Glaesserella parasuis	116	24
Haemophilus haemolyticus	68	4
Haemophilus influenzae	779	73
Haemophilus parainfluenzae	99	16
Influenza A virus	130	127
Mannheimia haemolytica	196	85
Pasteurella multocida	291	81
Salmonella enterica	12336	1066
Streptococcus suis	1623	72

E. coli, Influenza A virus, S. enterica, and S. suis are not in the Pasteurellaceae family; therefore, only their complete genomes in the nr/nt database were used for in silico marker specificity test. Species with [Actinobacillus] are not officially included in the Actinobacillus genus, but have not yet been assigned to a new genus (25).

TABLE 3

A. pleuropneumoniae-specific DNA marker candidates identified in silico

No.	Target	Locus tag in L20 (CP000569)	Predicted function	Length (NTs)	Match to 11 complete A. pleuropneumoniae genomes		Match to incomplete A. pleuropneumoniae genomes
					% query cover	% identity	No. of matches in 23 incomplete A. pleuropneumoniae genomes	% query cover	% identity
1	apxIVA-1	APL_0998	Toxin	385	100	100	51(match more than 1 contig in a genome)	19-100	79.43-100
2	apxIVA-2	APL_0998	Toxin	125	100	100	19	38-100	96.8-100
3	apxIVA-3	APL_0998	Toxin	326	100	100	23	96-100	99.08-100
4	apxIVA-4	APL_0998	Toxin	315	100	100	23	100	100
5	apxIVA-5	APL_0998	Toxin	116	100	100	23	100	100
6	eamA	APL_1023	EamA family transporter; DMT family transporter	203	100	100	23	100	99.51−100
7	nusG	APL_1717	Transcription termination/ anti-termination protein	139	100	100	23	100	100
8	sppA	APL_1268	Signal peptide peptidase, protease IV	105	100	100	23	100	100
9	xerD	APL_1542	Site-specific tyrosine recombinase	149	100	100	22(absent in contigs of ATCC 33377)	100	100
10	ybbN	APL_0080	Cochaperone YbbN; putative thioredoxin-like protein	127	100	100	23	100	100
11	ycfL	APL_0125	YcfL family protein; putative periplasmic lipoprotein	101	100	100	23	100	100
12	ychJ	APL_1658	YchJ family protein, hypothetical protein, SEC-C motif containing	140	100	100	24(present twice in strain 4226)	100	99.29−100

Percent query cover and percent identity after performing MegaBLAST searches against the nr/nt or whole-genome sequence (WGS) databases are shown. No similarity between marker candidates and sequences from other species was found by MegaBLAST.

Molecular validation of novel A. pleuropneumoniae-specific markers.

As the apxIVA gene, whose sequence is unique to A. pleuropneumoniae, is a proven A. pleuropneumoniae-specific marker (7, 9, 10, 17), we did not perform PCR to validate the five marker candidates within the apxIVA gene. The presence of seven other marker candidates in A. pleuropneumoniae and other bacteria was examined by PCR using primers specific to each marker candidate and specific to A. pleuropneumoniae genomes in silico (Table 4). Genomic DNA from reference strains of A. pleuropneumoniae covering all 19 serovars, 108 A. pleuropneumoniae field isolates covering serovars 1, 2, 5, 12, 15, and nontypables, eight other Actinobacillus and [Actinobacillus] species, and seven other bacterial species was used as PCR template. For all seven marker candidates (i.e., those within the eamA, nusG, sppA, xerD, ybbN, ycfL, and ychJ genes), PCR amplicons of expected sizes were detected in all A. pleuropneumoniae strains and isolates but were absent in other species (Table 5). Representative gel electrophoresis results are also shown in supplemental material. As controls, two pairs of previously reported apxIVA primers (9) were tested and shown to be A. pleuropneumoniae-specific, as expected, recognizing all A. pleuropneumoniae strains and isolates tested (Table 5). These results indicate that the seven sequences within the eamA, nusG, sppA, xerD, ybbN, ycfL, and ychJ genes are validated as novel A. pleuropneumoniae-specific DNA markers, can serve as additional or alternative targets for A. pleuropneumoniae detection assays, and are interchangeable with apxIVA. The use of more than one marker can prevent diagnostic evasion.

TABLE 4

Primers used in this study

Primer no.	Primer name	Sequence (5′ to 3′)
P228	eamA-F	CACTTCAAGTCGGCACTGTC
P229	eamA-R	TCATAATAATTGCAGCGTTAGTGA
P230	sppA-F	CCAACGACGTAAAGCGAATAA
P231	sppA-R	CGAACAGACTATCGTCGCT
P240	xerD-F	ATAACGTATCTAAAAACTGTTCG
P241	xerD-R	TAGAATATCTAGGAATAAAAGTAGC
P242	ychJ-F	CGGTTATTTTTTCAAAATTCTTTGC
P243	ychJ-R	CGCCTATTTAGCCTAATCC
P250	nusG-F	GGCTTTGTGATTTTATAAAATAAG
P251	nusG-R	GCCGATAAAAAACACTTTGTG
P254	ybbN -F	TCATTATTACGCCGGTTGGC
P255	ybbN -R	TCACGGTTGCCAATAAAAATTG
P256	ycfL-F	ACTCAACCAAGGTTGCATCG
P257	ycfL-R	AATCAAGGCATTACACAAACCAA
	ApxIVA-1L	TGGCACTGACGGTGATGA (9)
	ApxIVA-1R	GGCCATCGACTCAACCAT (9)
	ApxIVANEST-1L	GGGGACGTAACTCGGTGATT (9)
	ApxIVANEST-1R	GCTCACCAACGTTTGCTCAT (9)

TABLE 5

Validation of A. pleuropneumoniae-specific markers by PCR

Species	Serovar	Strain	No. of strains tested	apxIVA		Marker candidate
				1L-1R(422)	NEST 1L-1R(377)	eamA(192)	nusG(117)	sppA(83)	xerD(74)	ybbN(58)	ycfL(54)	ychJ(66)
A. pleuropneumoniae	1	ATCC 27088T,2 field isolates	3	++	++	++	++	++	++	++	++	++
	2	ATCC 27089,1 field isolate	2	++	++	++	++	++	++	++	++	++
	3	ATCC 27090	1	++	++	++	++	++	++	++	++	++
	4	ATCC 33378	1	++	++	++	++	++	++	++	++	++
	5	ATCC 33377,L20,ATCC 55454,100 field isolates	103	++	++	++	++	++	++	++	++	++
	6	ATCC 33590	1	++	++	++	++	++	++	++	++	++
	7	WF83	1	++	++	++	++	++	++	++	++	++
	8	405	1	++	++	++	++	++	++	++	++	++
	9	CVJ13261	1	++	++	++	++	++	++	++	++	++
	10	D13039	1	++	++	++	++	++	++	++	++	++
	11	56153	1	++	++	++	++	++	++	++	++	++
	12	8328,1 field isolate	2	++	++	++	++	++	++	++	++	++
	13	N-273	1	++	++	++	++	++	++	++	++	++
	14	3906	1	++	++	++	++	++	++	++	++	++
	15	HS143,1 field isolate	2	++	++	++	++	++	++	++	++	++
	16	A-85/14	1	++	++	++	++	++	++	++	++	++
	17	16287-1	1	++	++	++	++	++	++	++	++	++
	18	7311555	1	++	++	++	++	++	++	++	++	++
	19	7213384-1	1	++	++	++	++	++	++	++	++	++
	Nontypable	3 field isolates	3	++	++	++	++	++	++	++	++	++
A. equuli		ATCC 9346	1	-	-	-	-	-	-	-	-	-
[A.] indolicus		CCUG 39029T	1	-	-	-	-	-	-	-	-	-
A. lignieresii		ATCC 13372,CCUG 41384T	2	-	-	-	-	-	-	-	-	-
[A.] minor		CCUG 38923T	1	-	-	-	-	-	-	-	-	-
[A.] porcinus		CCUG 38924T	1	-	-	-	-	-	-	-	-	-
[A.] rossi		ATCC 27072	1	-	-	-	-	-	-	-	-	-
A. suis		ATCC 15557,ATCC 33415T	2	-	-	-	-	-	-	-	-	-
A. ureae		ATCC 25976	1	-	-	-	-	-	-	-	-	-
B. trehalosi		ATCC 33367	1	-	-	-	-	-	-	-	-	-
G. parasuis		ATCC 19417	1	-	-	-	-	-	-	-	-	-
		Field isolates	6	-	-	-	-	-	-	-	-	-
H. influenzae		ATCC 33391	1	-	-	-	-	-	-	-	-	-
M. haemolytica		ATCC 29696	1	-	-	-	-	-	-	-	-	-
P. multocida		ATCC 43137	1	-	-	-	-	-	-	-	-	-
		ATCC BAA-1113	1	-	-	-	-	-	-	-	-	-
S. Choleraesuis		ATCC 7001	1	-	-	-	-	-	-	-	-	-
S. suis		ATCC 43765	1	-	-	-	-	-	-	-	-	-

++, PCR product of expected size was present; -, no PCR product present; numbers in parentheses are expected PCR product sizes in base pairs (bp). Genomic DNA of various bacterial species/strains was tested for the presence of candidate marker sequences using PCR.

DISCUSSION

Identifying species-specific markers for A. pleuropneumoniae previously involved individually testing DNA fragments in cross-hybridization or PCR experiments (9, 13, 15, 20). These methods are time-consuming and incomprehensive, as only a limited number of DNA fragments can be tested. Moreover, some of the resulting detection assays are not species-specific and still show cross-reactivity with closely related species (3, 13, 15). Using comparative genome analysis, based on a strict criterion of 100% nucleotide identity across sequences of 100–400 nucleotides conserved in only 11 complete A. pleuropneumoniae genomes, 12 sequences were identified as putatively A. pleuropneumoniae-specific (Table 3). Other highly conserved sequences with less than 100% conservation among the 11 complete genomes were not considered here but may be useful as A. pleuropneumoniae-specific markers and require further investigation.

Even though 11 complete A. pleuropneumoniae genome assemblies covering serovars 1–5 and 7–8 (Table 1) were used for the initial step of A. pleuropneumoniae-conserved sequences identification, the A. pleuropneumoniae-conserved sequences were later tested for their A. pleuropneumoniae-specificity using the nr/nt nucleotide collection database and the WGS database limited to the Pasteurellaceae family, which include 34 complete and incomplete A. pleuropneumoniae genome assemblies covering serovars 1–13, 30 genome assemblies from 13 other Actinobacillus and [Actinobacillus] species, 116 G. parasuis genome assemblies, and 291 P. multocida genome assemblies (Table 2). Since large genome databases can be accessed and utilized, in silico genome analysis is a powerful tool to guide marker discovery. The more genomes of target species and closely related nontarget species become available, the higher accuracy and specificity of in silico marker discovery will be. Molecular validation is still necessary, especially for marker discovery of species with limited genome data. The more bacterial species and isolates that are available for molecular validation, the more accurate and specific the resulting markers will be.

Five marker candidates identified in this study are within apxIVA, previously reported to be an A. pleuropneumoniae-specific gene (9, 20), confirming that our in silico marker identification method is effective. Nonetheless, the five apxIVA sequences (apxIVA-1-5) identified in this study are not identical to those previously described. As there have been reports of atypical A. pleuropneumoniae isolates failing to amplify the predicted target with existing apxIVA-specific primers (5, 21), our new apxIVA targets provide alternative options for molecular confirmation of A. pleuropneumoniae.

Two apxIVA regions (apxIVA-1 and 2) identified herein are in the 3′ part of apxIVA and are in close proximity to the A. pleuropneumoniae-specific region previously identified in hybridization experiments and some previously published primer pairs (Fig. 1A) (7, 9, 20). The 5′ and central parts of the apxIVA gene were originally disregarded as A. pleuropneumoniae-specific because probes from these regions showed weak hybridization signals with A. lignieresii (9, 20). However, later studies identified additional conserved regions within the 5′ (17) and the central part of apxIVA (10) that can be used as targets for A. pleuropneumoniae molecular detection assays (Fig. 1A). Three newly identified marker candidates (apxIVA-3, 4, and 5) are within the central part of apxIVA (Fig. 1), but do not overlap the conserved regions previously reported (10, 17), as these sequences do not match our criteria of being 100% conserved among the 11 complete genomes. These combined results indicate that our marker discovery strategy does not identify all possible markers but is useful for identifying multiple effective species-specific markers simultaneously.

FIG 1

Locations of previously published primer pairs and newly identified apxIVA marker candidates in the genome of the A. pleuropneumoniae serovar 2 strain CCUG 47657 that contains only the apxIVA gene (A), and in the genome of the A. pleuropneumoniae serovar 8 strain MIDG2331 that contains both apxIVA and apxIV-S genes (B). Previously published primer pairs are shown as arrowheads. Green arrowheads denote primers oAPXIVA-TSP1 and oAPXIVA-TSP2 (17). Blue arrowheads denote primers apxIVA-exo-F and apxIVA-exo-R (10). Purple arrowheads denote primers ApxIVA-1L and ApxIVA-1R (9). Black arrowheads denote primers named apxIVA1 and apxIVA3 (7). Gray rectangles represent regions apxIVA-1 to apxIVA-5 identified in this study (Table 3). Region apxIVA-1′ is 90% identical to apxIVA-1. Region apxIVA-2′ is 94 to 98% identical to region 2. Green rectangles indicate homologous regions between apxIVA and apxIV-S.

In addition to apxIVA, some A. pleuropneumoniae strains also contain apxIV-S, a partial duplication of apxIVA that shares approximately 87% identity with apxIVA in the 3′ region (Fig. 1B) (22). In A. pleuropneumoniae genomes with both apxIVA and apxIV-S, the five new apxIVA marker candidates match to different regions but are still A. pleuropneumoniae-specific in silico (Fig. 1B, Table 3). Regions apxIVA-1′ and 2′ with 90% and 94–98% identity to apxIVA-1 and apxIVA-2, respectively, are also present (Fig. 1). Coamplification of apxIVA-1′ and apxIVA-2′ along with apxIVA-1 and apxIVA-2 is possible but does not alter PCR product sizes and thus detection results. The presence of multiple highly homologous regions in one genomic DNA molecule may serve as more targets for PCR, possibly leading to detection assays with higher sensitivity.

Although not encoding an intact ApxIV protein (NCBI accession no. NZ_LR134169), the NCTC 10568 A. lignieresii genome contains sequences (comprising multiple open reading frames) sharing 73% identity over 71% of the A. pleuropneumoniae apxIVA sequence (71% query cover), as determined by BLASTn. Five A. pleuropneumoniae-specific apxIVA marker candidates identified here do not share significant similarity with the apxIVA-like sequences in A. lignieresii, as determined by default parameters of MegaBLAST search against databases which include three complete and incomplete A. lignieresii genomes (Table 3). In short, five A. pleuropneumoniae-specific apxIVA regions are A. pleuropneumoniae-specific despite the presence of apxIVA-like sequences in A. lignieresii. Nonetheless, cross-reactivity with A. lignieresii in pig-derived samples is unlikely, as A. lignieresii is a pathogen of cattle and sheep (23).

In addition to sequences within apxIVA, seven novel marker candidates that map to various genes were identified. Six newly identified A. pleuropneumoniae-specific markers, namely, eamA, nusG, sppA, ybbN, ycfL, and ychJ, share 100% identity among all 11 complete A. pleuropneumoniae genome assemblies and 99.29–100% identity among all 23 incomplete A. pleuropneumoniae genome assemblies, confirming their highly conserved nature among A. pleuropneumoniae genomes. These six sequences are also A. pleuropneumoniae-specific compared in silico with available databases (Table 3) and when tested by PCR with DNA from available bacterial species and strains (Table 5).

The last marker candidate, xerD, shares 100% identity among all 11 complete A. pleuropneumoniae genomes but is found only in 22 out of 23 incomplete A. pleuropneumoniae genomes. The xerD marker candidate is absent in genome contigs of the ATCC 33377 strain (CABEFA01), suggesting that the ATCC 33377 genome may not contain xerD or the contigs that contain whole xerD marker sequence are absent in the genome assemblies. The xerD sequence identified is only 149 nucleotides in length. Assembling contigs to contain this short sequence should not be difficult unless the genome contains multiple sequences homologous to xerD. As seen in the case of apxIVA-2, when performing MegaBLAST searches against the Pasteurellaceae WGS database, only 19 out of 23 matches with A. pleuropneumoniae incomplete genomes (38–100% query cover and 96.8–100% identity) were observed (Table 3). Nonetheless, xerD-specific PCR product was observed when genomic DNA from the ATCC 33377 strain was used as the template (Table 5 and 6), indicating that xerD can also serve as a marker for A. pleuropneumoniae identification.

TABLE 6

Bacteria used in this study

Genus and species	Serovar	Strain name	Source/referencec
Actinobacillus pleuropneumoniae	1	ATCC 27088T	ATCC (33)
	2	ATCC 27089	ATCC (33)
	3	ATCC 27090	ATCC (33)
	4	ATCC 33378	ATCC (34)
	5a	ATCC 33377	ATCC (34, 35)
	5b	L20	(34, 35)
	5	ATCC 55454	ATCC
	6	ATCC 33590	ATCC (36)
	7	WF83	(37)
	8	405	(38)
	9	CVJ13261	(39)
	10	D13039	(40)
	11	56153	(41)
	12	8328	Denmark
	13	N-273	(42)
	14	3906	(43)
	15	HS143	(44)
	16	A-85/14	(45)
	17	16287-1	(46)
	18	7311555	(46)
	19	7213384-1	(5)
	1 [2]b2 [1]5 [100]12 [1]15 [1]Nontypable [3]	Field isolates from Thailand [108]	This study
Actinobacillus equuli		ATCC 9346	ATCC
[Actinobacillus] indolicus		CCUG 39029T	CCUG
Actinobacillus lignieresii		ATCC 13372	ATCC
		CCUG 41384T	CCUG
[Actinobacillus] minor		CCUG 38923T	CCUG
[Actinobacillus] porcinus		CCUG 38924T	CCUG
[Actinobacillus] rossi		ATCC 27072	ATCC
Actinobacillus suis		ATCC 15557	ATCC
		ATCC 33415T	ATCC
Actinobacillus ureae		ATCC 25976	ATCC
Bibersteinia trehalosi		ATCC 33367	ATCC
Glaesserella parasuis		ATCC 19417	ATCC
		Field isolates from Thailand [6]	This study
Haemophilus influenzae		ATCC 33391	ATCC
Mannheimia haemolytica		ATCC 29696	ATCC
Pasteurella multocida		ATCC 43137	ATCC
		ATCC BAA-1113	ATCC
Salmonella enterica subsp. enterica	Choleraesuis	ATCC 7001	ATCC
Streptococcus suis		ATCC 43765	ATCC

ATCC, american type culture collection; CCUG, culture collection university of gothenburg.

Numbers in brackets indicate the number of isolates. T indicates type strain of the species. Species with [Actinobacillus] are not officially included in the Actinobacillus genus, but have not yet been assigned to a new genus (25).

The Langford laboratory was the source of bacteria (or gDNA) that were not purchased from ATCC or CCUG. The growth and preparation of derived gDNA from these strains was carried out as described previously (5).

The use of multiple targets in a diagnostic assay can reduce false-negative results among A. pleuropneumoniae strains that may evade current detection methods. These novel A. pleuropneumoniae-specific markers could serve as targets for other DNA amplification assays such as isothermal amplification assays, which are more field-ready than PCR.

In conclusion, this study demonstrates how comparative genomics and molecular validation can accelerate species-specific marker discovery, save time, labor, and cost, and result in more markers compared with traditional marker discovery by hybridization or PCR experiments. The marker discovery strategy described herein can be applied to other species with sufficient genome data, leading to novel markers and diagnostic assays for infectious diseases.

MATERIALS AND METHODS

The experiments using Actinobacillus and other bacterial species were approved by BIOTEC and Chulalongkorn University Institutional Review Boards on Biosafety and Biosecurity with approval numbers BT-IBC-61-026 and IBC1831058, respectively.

A. pleuropneumoniae isolation from clinical samples.

A. pleuropneumoniae was isolated from lung or pleural fluid samples from pigs with clinical signs of respiratory disease submitted to the Livestock Animal Hospital, Chulalongkorn University, Nakhon Pathom, Thailand during 2017–2018, as per standard techniques (24). Briefly, clinical samples were cultured on blood agar (containing 5% sheep red blood cells) with a Staphylococcus aureus nurse streak and incubated at 37°C with 5% CO2. Hemolytic colonies with a satellite characteristic around the S. aureus streak were further tested by Gram staining, Christie–Atkins–Munch-Peterson (CAMP) reaction with S. aureus, and catalase and oxidase tests. Species validation and molecular serotyping were performed using multiplex PCR targeting apxIVA and cps genes (7).

Bacterial strains and growth conditions.

Bacterial strains used to test the presence of DNA markers in this study are either in the Pasteurellaceae family or are present in pigs as commensal or pathogenic bacteria (Table 6). Bacteria (or genomic DNA) were purchased from the American Type Culture Collection (ATCC) or Culture Collection of University of Gothenburg (CCUG) or obtained from the Langford laboratory as indicated (Table 6). Some [Actinobacillus] species such as [A.] indolicus, [A.] minor, and [A.] porcinus are not officially included in the Actinobacillus genus but have not yet been assigned to a new genus (25). These species are herein described as [Actinobacillus.] Actinobacillus and [Actinobacillus] species, Glaesserella parasuis, Pasteurella multocida, and Haemophilus influenzae were grown on chocolate blood agar supplemented with IsoVitalex (BBL, BD, Franklin Lakes, NJ, USA) at 37°C with 5% CO2. Bibersteinia trehalosi, Mannheimia haemolytica, Salmonella enterica serovar Choleraesuis, and Streptococcus suis were grown on brain heart infusion (BHI) plates at 37°C with 5% CO2.

In silico DNA marker identification.

In the initial step, 11 complete genome assemblies covering serovars 1–5 and 7–8 (Table 1) were selected for analysis in consideration for algorithm efficiency. Sequences of 100 – 400 nucleotides in length that share 100% identity among the 11 complete genomes were selected by a custom script as “A. pleuropneumoniae-conserved sequences.” In the second step, these A. pleuropneumoniae-conserved sequences were used as queries to search for highly similar sequences using MegaBLAST (26–28). Searches were performed against the nucleotide collection (nr/nt) database, which contains sequences from GenBank, EMBL, DDBJ, PDB, and RefSeq, but excludes draft whole-genome contigs (WGS). Nineteen A. pleuropneumoniae-conserved sequences were identified to be specific to A. pleuropneumoniae genomes compared with the nr/nt database. In the third step, these 19 A. pleuropneumoniae-conserved sequences were used as queries to search for highly similar sequences in the WGS database containing draft genome contigs, limited to sequences of the Pasteurellaceae family, using MegaBLAST. Twelve A. pleuropneumoniae-conserved sequences remained specific to A. pleuropneumoniae genomes in silico compared with the WGS database and were considered “A. pleuropneumoniae-specific marker candidates.” The number of genome assemblies of selected species (in the same family as A. pleuropneumoniae or also present in pigs) available for in silico comparison is shown in Table 2.

In silico primer design.

BLASTn (26, 27), suitable for identification of more dissimilar sequences, was used to identify sequences of non-A. pleuropneumoniae species that share more than 70% identity with A. pleuropneumoniae-specific marker candidates from the nucleotide collection (nr/nt) database. Multiple alignment of A. pleuropneumoniae-specific marker candidates and similar sequences from other species was performed using Clustal Omega (29, 30). Regions with high mismatch between A. pleuropneumoniae and non-A. pleuropneumoniae species were selected for PCR primer design. Primer BLAST (31) searches against the nr/nt database were used to confirm that the newly designed PCR primers (Table 4) yielded PCR products of expected size only when A. pleuropneumoniae genomes were used as template.

Genomic DNA purification and PCR amplification.

Genomic DNA of various bacterial species was extracted using a standard DNA purification protocol (32). PCR was performed using Taq DNA polymerase with Standard Taq Buffer (M0273, New England Biolabs, Ipswich, MA, USA) according to the manufacturer’s protocol. Briefly, PCRs were prepared to contain final concentrations of 200 μM dNTPs, 0.2 μM each primer (Table 4), 0.025 U/μl Taq DNA polymerase, and 1 ng/μL of bacterial genomic DNA. Thirty cycles of 95°C for 30 s, 60°C for 1 min, and 68°C for 1 min were performed using a C1000 Touch PCR Thermal Cycler (Bio-Rad, Hercules, CA, USA). PCR products were visualized by agarose gel electrophoresis followed by ethidium bromide staining. Alternatively, Luna qPCR Master Mix (M3003, New England Biolabs) was used according to the manufacturer’s protocol. Briefly, qPCRs were prepared to contain final concentrations of 0.25 μM each primer and 1 ng/μL of bacterial genomic DNA. Forty-five cycles of 95°C for 15 s and 60°C for 30 s were performed using a Bio-Rad CFX96 real-time PCR machine. Fluorescence signals indicative of the presence of PCR products were measured.