Of the 39 dogs sampled, 20 were male and 19 were female (Table 1). Most of the dogs were entire. The majority of dogs were adults (n = 25), with equal numbers (n = 7) of puppies and juvenile dogs.

The number (and percentages, in parentheses) of dogs infected with each haemoparasite is recorded.

^§ The Ti Tree results are the sum of those from Nturiya and Pmara.

*Six dogs were co-infected with both M. haemocanis and ‘Ca. M. haematoparvum’.

^†The novel haemoplasma was detected in two dogs: one from Nturiya (Ti Tree) and one from Bidyadanga.

^#Data available for 10 of the 11 dogs.

PCR results were available for all samples for all assays. DNA was successfully purified and amplified from all canine blood samples as determined by the presence of adequate levels of canine internal control gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH; threshold cycle: 18.2 to 23.3).

See Table 1 for a complete breakdown of results for each region (See Additional file 1: Table S1 for results of individual dog). Of the 39 dogs, 17 (44%) were infected, as determined by a positive quantitative PCR (qPCR) result, with M. haemocanis (threshold cycle: 22.3 to 38.2) and eight (21%) with ‘Ca. M. haematoparvum’ (threshold cycle: 27.2 to 35.2). Twenty dogs (51%) were infected with A. platys (threshold cycle: 24.4 to 41.3) as determined by qPCR, 16 of these dogs were also positive using the Anaplasmataceae family conventional PCR (cPCR). No samples were positive by the Anaplasmataceae family cPCR and negative by the A. platys qPCR. Amplicons from two positive samples using the Anaplasmataceae family cPCR assay were confirmed by sequencing to contain A. platys DNA. Seventeen (44%) of the 39 dogs were positive, as determined by cPCR, for a Babesia/Theileria spp. piroplasm. Amplicon sequencing of 15 of the 17 positive results indicated the presence of B. vogeli (insufficient amplification of target DNA from the remaining two positive samples resulted in a failure to sequence).

All dogs positive for M. haemocanis and/or ‘Ca. M. haematoparvum’ were also positive by the pan-haemoplasma assay, but two additional dogs were positive only on the pan-haemoplasma assay (threshold cycle: 36.7 & 37.9), corresponding to approximately one to ten haemoplasmas per reaction (equivalent to 5 μl DNA template); one from Nturiya (Ti Tree) and one from Bidyadanga. Of these dogs, one was negative for all other haemoparasites, and the other was positive for A. platys and B. vogeli. Sequencing of a 600 bp 16 S rRNA gene fragment amplified from these discordant samples showed that they were both infected with the same novel haemoplasma species [EMBL:HE577612] (Figure 2: phylogenetic tree). The novel haemoplasma had highest identity (82.8 to 84.9%) to the haemofelis group of haemoplasmas 4 , closely followed by the haemominutum group of haemoplasmas (79.1 to 83.7%) and the non-haemotropic mycoplasmas Mycoplasma fastidiosum and Mycoplasma cavipharyngis (82.3 and 82.5% respectively). The 23 bp deletion common to haemofelis group haemoplasmas was not present in the novel haemoplasma gene sequence. Attempts to amplify a larger 16 S rRNA gene fragment and a ribonuclease P ribosomal gene fragment from these two samples were unsuccessful.

Serological results were available for all dogs, except for an A. platys antibody result for the only dog (from Pmara, Ti Tree) that was positive for E. canis antibodies. Nine dogs (23.7%) were positive for A. platys antibodies (Table 1). Both dogs that were serologically positive for D. immitis antigen were from Nturiya. No dogs were positive for B. burgdorferi antibodies.

Six of the dogs in Ti Tree were infected with both M. haemocanis and ‘Ca. M. haematoparvum’, whilst co-infection with M. haemocanis and ‘Ca. M. haematoparvum’ was not present in the other communities. Six of the haemoplasma positive dogs were infected with A. platys as determined by qPCR: one was infected with M. haemocanis alone; two with ‘Ca. M. haematoparvum’ alone; two co-infected with M. haemocanis and ‘Ca. M. haematoparvum’; and a further dog infected with the novel haemoplasma. Six dogs positive for B. vogeli, as determined by cPCR, were positive for one or more haemoplasma: three with M. haemocanis alone; one with ‘Ca. M. haematoparvum’ alone; one co-infected with M. haemocanis and ‘Ca. M. haematoparvum’; and a further dog with the novel haemoplasma. Three dogs infected with M. haemocanis, one of which was co-infected with ‘Ca. M. haematoparvum’, were seropositive for A. platys. One dog co-infected with M. haemocanis and ‘Ca. M. haematoparvum’ was also seropositive for D. immitis antigen.

Of the 20 dogs infected with A. platys as determined by qPCR, 13 were cPCR-positive for B. vogeli, six were seropositive for A. platys antibodies (one dog had an unknown result), one was seropositive for E. canis antibodies, and one was seropositive for D. immitis antigen.

Of the 17 dogs infected with B. vogeli as determined by cPCR, five were seropositive for A. platys antibodies (one unknown), one was serpositive for E. canis antibodies, and none were seropositive for D. immitis antigen.

The dog that was seropositive for E. canis antibodies was also PCR positive for A. platys, B. vogeli, M. haemocanis and ‘Ca. M. haematoparvum’.

Due to the low numbers of dogs from each community, only haemoparasite prevalences from Central Australia (Ti Tree) and Western Australia (Bidyadanga) could be compared. Haemoplasma prevalence was highest (16/22; 73%) in Central Australia and lowest (2/9; 22%) in Western Australia. The difference in haemoplasma prevalence between these two sites was statistically significant (p = 0.017; Fisher's exact test). In contrast, B. vogeli prevalence was low in Central Australia (4/22; 18%) but higher (6/9; 66%) in Western Australia. The difference in B. vogeli prevalence between these two sites was statistically significant (p = 0.003; Fisher's exact test). There was a trend towards a statistically significant difference in A. platys prevalence between Central Australia (8/22; 36%) and Western Australia (7/9; 78%; p = 0.054; Fisher’s exact test).

EDTA-blood and whole blood (clot tube) were collected from 39 free-roaming dogs associated with four Aboriginal communities, between February 2008 and October 2009. Blood was stored at 4°C, and transported to the laboratory within seven days of collection. Serum was harvested from whole blood after centrifugation. EDTA blood and serum were subsequently frozen at −70°C until required.

Most dogs were of mixed breed, and were considered variable hybrids of the dingo (Canis lupus dingo) and the domestic dog (Canis lupus familiaris). Based on dentition they were assigned to three age classes: puppy (< 3 months), juvenile (3 to 12 months) and adult (> 12 months).

DNA was purified from the samples using the QIAmp® DNA Mini Kit (Qiagen Pty Ltd) according to the manufacturer’s recommendations, then shipped to the UK on dry ice.

Purified DNA was subjected to an Anaplasmataceae family conventional PCR cPCR and a Babesia / Theileria spp. cPCR. The Anaplasmataceae family cPCR comprised a primer pair (EHR16SD GGTACCYACAGAAGAAGTCC & EHR16SR TAGCACTCATCGTTTACAGC) 14 , which amplifies a 345 base pair (bp) fragment from the 16 S rRNA gene using the following thermal protocol: 95°C for 10 min and 40 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s. The Babesia/Theileria spp. cPCR comprised a primer pair (BabgenF GAAACTGCGAATGGCTCATTA & BabgenR CGGTAGGCCAATACCCTACCGTC) 15 , which amplifies a 250–270 bp fragment from the 18 S rRNA gene using the following thermal protocol: 94°C for 10 min and 40 cycles of 94°C for 30 s, 65°C for 30 s, and 72°C for 45 s. All reactions used 2x HotStarTaq Master Mix (Qiagen, Crawley, UK) with 200 nM each primer pair, 3.0 mM final MgCl₂ and 2 μL DNA template in a total volume of 20 μL, and were performed in a Tetrad thermal cycler (MJ Research, Waltham, MA, USA). Positive and negative control reactions were used in each assay. Amplification products were visualised on an agarose gel following electrophoretic separation.

Purified DNA was also subjected to the following qPCR assays: species-specific qPCRs for each of M. haemocanis and ‘Ca. M. haematoparvum’, each duplexed with a GAPDH qPCR 3 , pan-haemoplasma qPCRs 16 , and an A. platys-specific qPCR (personnel communication: M Robinson, Acarus Laboratory, University of Bristol). Briefly, the canine species-specific qPCRs were performed using 2x HotStarTaq Master Mix with 200 nM species-specific haemoplasma 16 S rRNA gene primer pair (M. haemocanis: Mhf 1167 F GTGCTACAATGGCGAACACA & Mhf 1246R TCCTATCCGAACTGAGACGAA; ‘Ca. M. haematoparvum’: CMhp 124 F GGAGAATAGCAATCCGAAAGG & CMhp 252R GCATTTACCCCACCAACAAC), 100 nM haemoplasma TaqMan probe (M. haemocanis: Mhf 1188 T FAM-TGTGTTGCAAACCAGCGATGGT-BHQ1; ‘Ca. M. haematoparvum’: CMhp 192 T FAM-CTTCGGGAGCCCCGCGC-BHQ1), 25 nM canine GAPDH gene primer pair (17 F TCAACGGATTTGGCCGTATTGG & 106R TGAAGGGGTCATTGATGGCG), 50 nM canine GAPDH TaqMan probe (TXR-CAGGGCTGCTTTTAACTCTGGCAAAGTGGA-BHQ2), 4.5 mM final MgCl₂ and 5 μl gDNA in a total volume of 25 μl. Briefly, the pan-haemoplasma qPCRs were performed using 2x HotStarTaq Master Mix with 200 nM haemoplasma 16 S rRNA gene primer pair (Haemofelis group: HF grp 567 F GGAGCGGTGGAATGTGTAG & HF grp 680R GGGGTATCTAATCCCATTTGC; Haemominutum group: HM grp 1061 F GGGGCCAAGTCAAGTCATC & HM grp 1199R GCGAATTGCAGCCTTTTATC), 100 nM haemoplasma TaqMan probe (Haemofelis group: HF grp 595P FAM-TYAAGAACACCAGAGGCGAAGGCG-BHQ1; Haemominutum group: HM grp 1096P FAM-TACCATTGTAGCACGTTYGCAGCCC-BHQ1), 4.5 mM final MgCl₂ and 5 μl gDNA in a total volume of 25 μl. All haemoplasma qPCRs were performed in an Agilent MX3005P real-time PCR machine (Agilent Technologies UK Ltd., Wokingham, UK): 95°C for 15 min and 45 cycles of 95°C for 10 s and 60°C for 30 s, during which fluorescence data were collected. Briefly, the A. platys-specific qPCRs were performed using 2x HotStarTaq Master Mix (Qiagen, Crawley, UK) with 200 nM A. platys-specific citrate synthase gene (gltA) primer pair (APLgltA1.f AGGCGTGATTTCATCCTTCA & APLgltA1.r CACAGCAAGCTCTTCATTTCC), 100 nM gltA TaqMan probe (APLgltA1.p FAM-TGGCTGCGAAGTATCATGGGGA-BHQ1), 5.0 mM final MgCl₂ and 5 μl gDNA in a total volume of 25 μl. All reactions were performed in an Opticon (Bio-Rad Labs. Ltd., Hemel Hempstead, UK) real-time PCR machine: 95°C for 15 min and 45 cycles of 95°C for 10 s and 64°C for 15 s, during which fluorescence data were collected. Positive and negative control reactions were used in each assay. Samples with discordant haemoplasma results (generic assay positive but M. haemocanis & ‘Ca. M. haematoparvum’ negative) were subjected to cPCR amplification using 2x HotStarTaq Master Mix, 200 nM of a universal Mycoplasma primer pair (HBT-F ATACGGCCCATATTCCTACG & HBT-R TGCTCCACCACTTGTTCA) 17 , 3.75 mM final MgCl₂ and 5 μL DNA template in a total volume of 25 μL, using the following thermal protocol: 95°C for 15 min and 50 cycles of 95°C for 10 s, 55°C for 15 s and 72°C for 30 s in a MJ Mini thermal cycler (Bio-Rad Labs. Ltd.).

Amplicons from positive cPCR results were purified using the NucleoSpin® Extract II Kit (Macherey-Nagel, ABgene, Epsom, UK), and subjected to DNA sequencing using Applied Biosystems Big-Dye Ver 3.1 chemistry on an Applied Biosystems model 3730 automated capillary DNA sequencer. BLASTn analysis 18 was performed to compare the 16 S rRNA gene (Anaplasmataceae family & haemoplasmas) and 18 S rRNA gene (Babesia/Theileria spp.) sequences obtained to those in GenBank. A phylogenetic tree including existing haemoplasma species, as well as selected non-haemoplasma Mycoplasma species, was constructed using MacVector version 12 (MacVector, Inc., Cambridge, United Kingdom) for the 16 S rRNA gene using the neighbour-joining program from a distance matrix 19 , corrected for nucleotide substitutions by the Kimura two-parameter model 20 . The data set was re-sampled 1000 times to generate bootstrap percentages. The 16 S rRNA gene fragment of the novel haemoplasma species was deposited in the European Molecular Biology Laboratory Nucleotide Database (HE577612).

Serum samples were tested by enzyme-linked immunosorbent assay (ELISA) for Dirofilaria immitis (heartworm) antigen and Ehrlichia canis, Borrelia burgdorferi, Anaplasma phagocytophilum antibodies using the SNAP 4Dx kit (Idexx Laboratories Pty. Ltd., Rydalmere, NSW 2116, Australia). Due to cross reactivity between A. platys- and A. phagocytophilum-directed antibodies, in conjunction with the absence of A. phagocytophilum from Australia, a positive A. phagocytophilum ELISA result was taken to indicate exposure to A. platys (Personal Communication: Idexx Laboratories Pty. Ltd.).

Data were entered into Microsoft® Office Excel® 2007 (Microsoft Corporation) and statistical analyses performed using Statistical Package for the Social Sciences version 18.0 (SPSS Inc., Chicago, IL). Categorical data, i.e. absence or presence of specific infection, gender, age, Aboriginal community (Ti Tree & Bidyadanga only), were compared using the chi-square test, or the Fisher’s exact test where sample sizes were small. Too few dogs were available from Tiwi Islands and Goodooga to allow statistical evaluation of area influence on infection. Insufficient numbers were available in each category to enable multivariate analysis. Significance was taken as p ≤ 0.05.

All procedures were performed with the approval of the Animal Ethics committee of the University of Sydney (N00/11-2006/3/4492).