Wolbachia endobacteria infect most species of insect and are present in other arthropod groups as well as in most filarial nematode species. Phylogenetic analyses currently indicate as many as eight distinct Wolbachia lineages, designated supergroups A to H, along with some other lineages whose taxonomic position remains unresolved 12. Arthropod Wolbachia are found in all supergroups except C and D, with the majority of insect Wolbachia strains in supergroups A and B. Wolbachia from filarial nematodes are exclusively in supergroups C and D with the exception of endosymbionts from Mansonella spp., which are in supergroup F along with the Wolbachia from certain termites.

The association between Wolbachia and filarial nematodes appears to be one of mutualism, probably of an obligatory nature 3. Elimination of the endosymbionts by antibiotic treatment disrupts embryogenesis and hence microfilarial production, disrupts growth and development, and leads to macrofilaricidal effects. Wolbachia are also implicated in the immunopathology of infected persons and may contribute to the inflammatory adverse events seen after standard anti-filarial chemotherapy 3. Thus, targeting the Wolbachia endosymbiont has emerged as an attractive new strategy for filarial disease control 3.

Until recently, despite extensive investigation of diverse nematode groups, there had been no identification of Wolbachia in any non-filariid nematode 45. However, Wolbachia ftsZ, wsp (Wolbachia surface protein) and 16S rDNA sequences were recently amplified by PCR from DNA preparations of the metastrongylid nematode, Angiostrongylus cantonensis 6. Based on phylogenetic analysis of the wsp sequence, the apparent endosymbiont from A. cantonensis appeared to have a lineage distinct from the filarial Wolbachia (supergroups C, D or F) and was tentatively positioned in supergroup G, containing the Wolbachia from certain spiders such as Diaea circumlita. Both A. cantonensis and A. costaricensis are occasional pathogens of humans, the former a leading cause of eosinophilic meningitis in Asia and Pacific Islands while the latter produces abdominal disease in the Americas 78. The unexpected detection of Wolbachia in A. cantonensis and the medical implications of targeting this endosymbiont for novel antibiotic therapies for control of eosinophilic meningitis prompted us to investigate the status of Wolbachia in the genus Angiostrongylus in more detail.

A. cantonensis was obtained from Prof. Kentaro Yoshimura and maintained as a laboratory life-cycle in the Department of Parasitology of the Akita University Medical School. A. costaricensis strain "Santa Rosa" has been maintained in the Laboratório de Parasitologia Molecular, Instituto de Pesquisas Biomédicas da PUCRS since 1992 (in mice and the wild rodent Oligoryzomis nigripes; veronicelid slugs and Biomphalaria glabrata snails). Adult Brugia malayi (TRS labs) were used as a positive control for PCR and immunohistochemistry

PCR analysis of Wolbachia genes from both A. cantonensis and A. costaricensis adult worms showed no evidence that either species harbours Wolbachia endosymbionts. Genomic DNA from individual adult female A. cantonensis and A. costaricensis, which had been stored in 80% ethanol, was isolated (QiaAmp^® DNA mini kit, Qiagen) and analysed for Wolbachia by PCR as previously described 6 with modifications. PCR was carried out on a iCycler thermocycler (Bio-Rad) using the following conditions: 95°C for 4 min, followed by 40 cycles of 94°C for 15 s, 48°C for 30 s, 72°C for 2 min, then 72°C for 10 min, using primer pairs widely used for detection of Wolbachia from diverse hosts, and previously used on DNA from A. cantonensis 6. The following primers were used for amplification of Wolbachia gene sequences: wsp (wsp81F; 5'-TGG TCC AAT AAG TGA TGA AGA AAC-3' and wsp691R; 5'-AAA AAT TAA ACG CTA CTC CA-3'); ftsZ (ftsZ357F; 5'-CAA AAA TAT GTG GAT ACG CTC ATT GT-3' and ftsZ788R; 5'-GTA GCA CCA AAT ATT ATA TTT GCA TTT TC-3'); and 16S rRNA (16SwolbF; 5'-GAA GAT AAT GAC GGT ACT CAC-3' and 16SwolbR3; 5'-GTC ACT GAT CCC ACT TTA AAT AAC-3'). PCR reactions were performed in 25 μl containing 1 μl gDNA, 0.3 μM of each primer, 0.2 mM dNTPs, 1.5–3.0 mM MgCl₂ and 0.625 U of Taq polymerase (New England Biolabs [NEB]) in 1× reaction buffer (NEB). Positive PCR amplification was shown using DNA isolated from individual adult female B. malayi, which are known to contain Wolbachia, thus demonstrating that the primers and conditions were optimal for Wolbachia detection. All DNA samples produced nematode specific gene products. Angiostrongylus 18S rRNA, based on GenBank sequences from A. cantonensis and A. costaricensis ([GenBank:AY295804] and [GenBank:EF514913], respectively) was amplified with primers Ac18S 30F; 5'-AAG TGA AAC TGC GAA CGG CT-3' and Ac18S 830R; 5'-TCA CCT CTC GCG CAG GGA TA-3', while B. malayi gst was amplified as previously described 9 using GST 1377; 5'-TGC TCG CAA ACA TAG TAA TAG T-3' and GST 1632; 5'-ATC ACG GAC GCC TTC ACA G-3', indicating that there was DNA at sufficient concentrations for detection by standard one-round PCR.

In order to detect Wolbachia by immunohistochemistry, A. cantonensis and A. costaricensis worms were fixed in 80% ethanol and embedded in paraffin blocks. Sections were stained using affinity purified anti-Wolbachia peptidoglycan-associated lipoprotein (WoLP) antibodies and rabbit polyclonal anti-sera raised to Wolbachia surface protein (WSP) and visualized using the UltraVision ONE detection system (Lab Vision, ThermoFisher Scientific) using haematoxylin as a counterstain. In contrast to the positive staining observed in sections of B. malayi, both Angiostrongylus species were found to be negative with each of the Wolbachia-specific antibodies or antisera. It should be noted that both of these reagents are able to detect Wolbachia in species as diverse as filarial nematodes and the mosquito Aedes albopictus [91011, M. Taylor, unpublished data], indicating that a lack of cross-reactivity to Wolbachia found in Angiostrongylus is unlikely.

Because we were unable to reproduce the detection of Wolbachia in Angiostrongylus spp by PCR or immunohistochemistry, we analyzed in detail the sequences deposited in the GenBank database as part of the earlier report 6. Comparison of the wsp nucleotide sequence [GenBank:AY508980] to the non-redundant nucleotide database (nr) using BLASTN revealed good identity (97%) to the wsp sequence [GenBank:AY486092] from the putative supergroup G Wolbachia from the spider, D. circumlita c2, as described previously 6. However, the best identity (99%) was to wsp from the Wolbachia of the mosquito, Malaya genurostris [GenBank:AY462865]. Wolbachia multilocus sequence typing (MLST) has shown that phylogenetic inference based upon wsp sequences yields spurious lineages due to the high levels of intragenic recombination in this gene 1. MLST of the Wolbachia of the spider, D. circumlita c2, indicates that this endosymbiont is more correctly a member of supergroup A 1, and so too presumably is the Wolbachia from the mosquito, M. genurostris.

We performed similar nucleotide comparisons of the 16S [GenBank:AY652762] and ftsZ [GenBank:DQ159068] sequences attributed to Wolbachia from A. cantonensis and constructed phylogenetic trees for each. Wolbachia sequences used for multiple alignments and phylogenetic tree construction are as follows, with the first GenBank sequence for each invertebrate host organism corresponding to 16S and the second to ftsZ:

B

mal

Ri

In conclusion, our inability to detect Wolbachia in two different species of Angiostrongylus by PCR or immunohistochemistry argues against the presence of this endosymbiont in these metastrongylid nematodes. Lateral gene transfers from Wolbachia to invertebrates are common 13. Such a phenomenon could conceivably have resulted in the presence of Wolbachia fragments in the A. cantonensis genome, but since we were unable to detect Wolbachia sequences using the same PCR primers as were used in the earlier report 6, this possibility seems most unlikely. Instead, we suspect contamination of the DNA samples or PCR reactions with Wolbachia DNA from the mosquito, M. genurostris in the case of wsp and the filarial nematode, D. immitis, in the case of both ftsZ and 16S. This seems very plausible since the nucleotide identity is 99% in all cases. In support of this conclusion, we note that the wsp sequence of the M. genurostris endosymbiont was deposited in the GenBank database by the same authors of the recent report on Wolbachia in A. cantonensis, and that they carry out research on D. immitis. We cannot rule out the possibility that A. cantonensis from Taiwan contain Wolbachia while the worms we analyzed from Japan do not. However, the high identities of the reported sequences to those from arthropod Wolbachia (supergroup A) on the one hand, but to the Wolbachia from the nematode, D. immitis (supergroup C), on the other, evoke either a double Wolbachia infection, a phenomenon never observed in nematode-Wolbachia symbioses, or a highly divergent Wolbachia lineage unlike any reported thus far. The most straightforward conclusion from our analysis is that Angiostrongylus sp. do not contain Wolbachia.

The authors declare that they have no competing interests.

MT and JF wrote the paper. JF and SK performed bioinformatic analyses and constructed the multiple sequence alignments and phylogenetic trees. LF carried out PCR analyses and KJ immunohistochemistry and both contributed to writing the paper. RB and C G-T performed the initial PCR analysis and provided parasite material. MT conceived and directed the study. All authors read and approved the final manuscript.