LONEPINELLA SP. ISOLATED FROM WOUND INFECTIONS OF KOALAS

Members of the bacterial family Pasteurellaceae are mainly present on the mucosal surfaces of the respiratory, alimentary, or genital tracts of many mammalian species (Christensen et al. 2020), including marsupials such as koalas (Phascolarctos cinereus; Osawa et al. 2020). Lonepinella koalarum is a member of the family which was first isolated from the fecal biota of healthy koalas at Lone Pine Sanctuary, Queensland, Australia (Osawa 1992), with no association with disease. The genus Lonepinella, with only one named species, Lonepinella koalarum, has been characterized as a facultatively anaerobic, Gram-negative, rod-shaped bacterium that is catalase negative (Osawa et al. 1995) with variable reports on the presence of cytochrome C oxidase (Osawa et al. 2020). Koalas are highly adapted to use the leaves of various Eucalyptus spp., known to have high levels of tannins (Callaghan et al. 2011). Lonepinella koalarum has been defined as symbiotic with the koala, helping to break down tannin protein complexes in the diet (Osawa et al. 1995).

In 2019, Lonepinella-like isolates were associated with human wound infections following koala bites (Sinclair et al. 2019). Additionally, 22 bacterial isolates from the oral cavity of apparently healthy koalas from wildlife centers in Australia were investigated by amplification, sequencing, and phylogenetic analysis of two housekeeping genes, rpoA and recN, with the majority not being able to be taxonomically assigned to any named species within the Pasteurellaceae family (Chong et al. 2020). Although 6 of the 22 koala isolates were clustered closely within the genus Lonepinella, none of them were identified as L. koalarum (Chong et al. 2020). The occurrence of Lonepinella-like organisms in the oral cavity of koalas kept at a wildlife center in Tasmania, Australia has been previously reported (Hansen et al. 2017).

The Pasteurellaceae family is in constant flux, with renaming of the existing genera and description of new species such as Glaesserella australis (Turni et al. 2020). Here we report two cases of koala wound infection associated with a potentially new species within the genus Lonepinella in this family.

Case 1 was seen in early 2018 and involved a captive koala held at Dreamworld, Australia, Queensland with a nailbed wound infection. This animal had a history of nailbed infections going back to July 2017 that were thought to be the result of trauma when the animal was jumping onto the mesh barrier of the enclosure in which she resided. Initially the koala was treated orally with ketoconazole (Nizoral, Janssen-Cilag, Macquarie Park, NSW, Australia; 10 mg/kg) once daily and combined trimethoprim (4 mg/kg) and sulfamethoxazole (25 mg/kg) twice daily for 3 wk (Tribrissen 20, Jurox, Rutherford, NSW, Australia). Intermittent local treatment of the affected digit, including debridement, drainage of the affected area, and nail trimming was instituted.

Case 2 was a wild, adult koala monitored as part of a koala population management program in southeast Queensland (center point –27°15′54.3″, 152°59′24.72″). He was captured in April 2021 because of low activity data detected by his biotelemetry device (K-Tracker telemetry system, LX Group, Sydney, New South Wales, Australia) and unusual behavior. Upon examination, bite wound injuries consistent with intermale fighting were detected, including subluxation of the distal phalanx and a nailbed wound infection of the first digit of the right foreleg, and superficial multifocal trauma in the oral cavity. The koala was admitted to Endeavour Veterinary Ecology clinic, Toorbul, Queensland, Australia, and initially treated with a subcutaneous injection of 15 mg/kg amoxycillin trihydrate (Betamox LA, Norbrook, Tullamarine, Victoria, Australia) and provided with analgesia: paracetamol (15 mg/kg per every 6–12 h; Panadol Children, GlaxoSmithKline, Abbotsford, Victoria, Australia) and carprofen (4 mg/kg subcutaneous injection on day one, 2 mg/kg every 24 h for the next 2 d; Rimadyl, Zoetis, Rhodes, New South Wales, Australia).

Swab samples from the wound infection of the two koalas were transferred to the laboratory for bacterial culture and direct Gram staining. The swabs were inoculated on Columbia agar with 5% horse blood (HBA) and chocolate agar (Edwards Group Pty Ltd., Murarrie, Queensland, Australia), and incubated at 35 C, in 5% CO₂ for 48 h. Potentially significant organisms were single-colony picked and assessed on colony morphology, Gram staining, and the capacity of organisms to produce cytochrome C oxidase, catalase, and indole. Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry using VITEK MS (bioMérieux, Carponne, France) was performed, as well as commercial identification kits RapID ANA II System (Remel, Dartford, UK) and API 20 NE Microbial Identification System (bioMérieux, Carponne, France).

The two isolates were subjected to whole genome sequencing (WGS). The DNeasy ultra-clean microbial kit (Qiagen Sciences, German-town, Maryland, USA) was used as per manufacturer's instructions to extract the genomic DNA from overnight culture. The quality and quantity of the extracted DNA was then tested via DNA NanoDrop Spectrophotometer (Thermo Fisher Scientific, Wilmington, Delaware, USA) and gel electrophoresis (BioRad, South Granville, New South Wales, Australia).

The DNA samples were sequenced on an Illumina NextSeq 500 platform (150–base pair [bp] paired ends) by the Australian Centre for Ecogenomics Sequencing Service at the University of Queensland (St. Lucia, Queensland, Australia). Trimmomatic version 0.39 (Bolger et al. 2014) was used for filtering the raw reads with the quality of the reads assessed before and after filtering using FastQC version 0.11.5 (Babraham Bioinformatics 2022). Shovill version 1.1.0 with implemented SPAdes version 3.15.3 (Bankevich et al. 2012) was used for de novo assemblies with minimum length and minimum coverage set at 250 bp and 12 times, respectively. CheckM version 1.0.13 (Parks et al. 2015) was used to check the completeness and contamination and quast version 5.0.2 (Gurevich et al. 2013) was used for further quality control of the de novo assemblies. ABRicate version 1.0.1 (Seemann 2022) was used to search the draft assemblies for presence of known antimicrobial resistance genes using databases resfinder and card. The draft genome assemblies were then classified using the de novo workflow in the Genome Taxonomy Database Toolkit (GTDB-Tk) version 1.7.0 (Chaumeil et al. 2020) within the Pasteurellaceae family (using the relevant flag –taxa_filter in gtdbtk align function). A phylogenetic tree was inferred using the FastTree version 2.1.10, implemented in the de novo workflow, with the 45,555 GTDB genomes from GTDB-Tk reference data v.r202. This included a total of 133 genomes from the family Pasteurellaceae plus the two assemblies from the current study and the draft genome assembly of L. koalarum strain UCDLQP1 (Dahlhausen et al. 2020), with genus Gallibacterium as the outgroup. The average nucleotide identity (ANI) between the genome of the isolates and that from L. koalarum–type strain was calculated using the ani_rep workflow in the GTDB-Tk. Isolates sharing an ANI similarity values of more than 95% were considered to be from the same species, and those with similarity values between 80 and 95% were considered to be from the same genus (Chun et al. 2018).

In addition, the recN sequence from the Lonepinella-like isolates previously obtained from a human wound infection (Sinclair et al. 2019) was used as the query fasta file for extraction of the recN gene sequences from the draft assemblies using BLASTn version 2.2.31+ as previously described (Omaleki et al. 2020). The sequence was then compared to those of the human wound infection, the Lonepinella koalarum type strain, ACM3666, as well as isolates previously obtained from the oral cavity of koalas (Table 1; Chong et al. 2020) with the recN from Actinobacillus porcinus NM319 as the outgroup. Geneious Prime version 2022.1.1 (Biomatters, Inc. 2022) was used for performing the alignments with implemented ClustalW version 2.1 (Larkin et al. 2007). Additionally, genome relatedness between the isolates was calculated using the recN sequence and the previously suggested formula (Zeigler 2003), with the results also compared to that from the ANI calculation of the draft assemblies. The software PhyML version 3.3.2 (Guindon et al. 2010) implemented in Geneious was used for phylogenetic analysis of recN gene alignments with a general-time reversible nucleotide substitution used for tree construction (bootstrap 1,000). The generated phylogenetic trees were visualized using FigTree version 1.4.2 (Rambaut 2018).

Although the initial lesion associated with the fourth digit of the right foreleg appeared to resolve, similar lesions appeared on the fourth digit of the left foreleg, while the lesion on the right foreleg recrudesced. Treatment with oral ketoconazole was reinstituted; however, a new lesion appeared on the fourth digit of the left hind leg. Despite medical and surgical treatment of the lesions, they failed to heal. The animal did not use the limbs normally and appeared to be in pain. As a result, all three distal phalanges were amputated at the P2/P3 joint in September 2018.

The oral wounds healed well, but the nailbed infection failed to improve, despite a second dose of amoxycillin trihydrate. Regular cleaning of the nailbed wound with 0.05% chlorhexidine gluconate, sterile saline flushing, and manuka honey dressings was instituted, alongside daily subcutaneous injections of 60 mg/kg chloramphenicol (Chloramphenicol 150, Ceva, Glenorie, New South Wales, Australia) for a total of 19 doses. Although this resulted in some improvement, the associated claw became increasingly loose and required removal. The wound and exposed hyponychium healed and the koala was observed to be using the affected digit comfortably in the enclosure. He was released after 34 d and his progress in the wild monitored closely. Upon recapture in October 2021, there was no evidence of recurrence of the infection and the claw had partially regrown.

The Gram-stained smear from both cases showed a moderate number of leucocytes and Gram-negative bacilli, with a few Gram-positive cocci and bacilli.

The bacteriological culture from Case 1 resulted in a light pure growth of a Gram-negative bacillus that could not be identified using routine diagnostic methods. The organism had the capacity to produce cytochrome C oxidase, but was negative for catalase and indole (isolate BR2882).

From Case 2, a light mixed growth of Citrobacter koseri and Enterobacter cloacae was identified. A third organism, a gram-negative bacillus, was also obtained from this case (BR2904) with similar characteristics to the one from Case 1 in 2018. In both cases, the VITEK MS was unable to identify the organism. The results from the two commercial kits were inconsistent, with API 20 NE finding the significant taxa as Mannheimia haemolytica/Bibersteinia trehalosi and RapID finding it as Capnocytophaga sp. The isolates were forwarded to the Microbiology Research Group at the University of Queensland.

The draft assembly of the two koala wound isolates, BR2882 and BR2904, contained 44 and 37 contigs respectively with total length of 2.19 Mbp and a GC content of 38.8 % (Supplementary Table S1). CheckM analysis identified the genomes as belonging to Gammaproteobacteria but could not assign them to any genus (Supplementary Table S1). The isolates shared an ANI of 92.8% with the L. koalarum type strain from the GenBank database while sharing a mean of 98% identity to each other.

Phylogenetic analysis of the draft genome assemblies using the GTDB platform clustered the two koala isolates closely to the L. koalarum type strain as well as to the previously isolated and sequenced isolate UCD_LQP1 obtained from the feces of a female healthy koala at the San Francisco Zoo, San Francisco, California, USA (Dahlhausen et al. 2020; Fig. 1, GenBank accession number GCF_008723255.1). The recN sequences from the two isolates were extracted from the draft assembles and compared to each other as well as to those from the previously published isolates from koala oral cavity (Chong et al. 2020) and human wound infection (Sinclair et al. 2019). The recN from BR2904 obtained from the wild koala was 100% identical to that from human wound infection isolate MS14434 while sharing 99.69% identity to that from isolate BR2882 (genome relatedness of 0.95 and 0.94, respectively). The two koala wound isolates shared 97.17% identity with L. koalarum type strain in their recN gene sequences, resulting in the calculated genome relatedness of 0.886 between them and the L. koalarum type strain. The maximum-likelihood analysis of the recN sequence placed the two koala wound isolates in close relationship with the previously obtained human wound isolate, MS14434 (Fig. 2).

The majority of reported skin infections in animals are associated with Staphylococcus spp. (Foster 2012). However, the association of members of the Pasteurellaceae family, especially Pasteurella multocida, with human wound infection following cat and dog bites has been long known (Hubbert and Rosen, 1970).

Previous studies have demonstrated the adequacy of recN gene analysis in assessing the genome relatedness of the members of the Pasteurellaceae family, with values more than 0.85 suggesting the compared isolates probably belong to the same species and values more than 0.4 suggesting that isolates are probably in the same genus (Zeigler 2003; Kuhnert and Korczak 2006). The calculated genome similarity index of 0.886 between the two wound isolates and that of the L. koalarum type strain from the recN gene falls towards the lower limit of the species threshold. However, the 92.8% ANI between the draft genome assemblies suggest that the two koala wound isolates do not belong to the species L. koalarum, but do belong to the genus Lonepinella. This result is also reflected in the GTDB phylogenetic tree where the two koala wound isolates form a separate cluster to that formed by the L. koalarum type strain and isolate UCD_LQP1. Hence our results from ANI calculation strongly suggest that although both isolates belong to the same species within the genus Lonepinella, they are not L. koalarum. Additionally, both isolates shared a high level of similarity in their recN gene to that from isolate MS14434 previously obtained from a human wound infection following a koala bite, a relationship also reflected in the phylogenetic tree constructed based on the recN sequences. This suggests that these three isolates form a potential new species within the genus Lonepinella. Importantly, this possible new species has the capacity to cause human and koala wound infection and is a potential zoonotic organism.

Previous studies have described a total of nine Lonepinella-like isolates obtained from the oral cavity of both captive koalas (six isolates) and wild koalas (three isolates) under care for different reasons, via culture (Hansen et al. 2017; Chong et al. 2020). These nine isolates were from geographically dispersed locations in Australia, including Queensland, New South Wales, and Tasmania. Combining the results of Chong et al. (2020) with the isolates in the current study and the isolates associated with human bite wounds (Sinclair et al. 2019), indicates that Lonepinella-like organisms are prevalent in the oral cavity of koala populations. However, recent studies on the koala fecal microbiome using culture-independent techniques have suggested a low abundance of L. koalarum in the koala gastrointestinal tract (Barker et al. 2013; Blyton et al. 2022). It would seem most likely that the infections seen in our two cases were due to oral contamination of tissue damaged by physical insult (Case 1 was associated with trauma from cage wire, and Case 2 had trauma that appeared typical of intermale fighting). As human wound infections following koala bites have been linked with Lonepinella-like organisms (Sinclair et al. 2019), it is not surprising that Lonepinella-like organisms may infect koala wounds associated with trauma events. These reports make clear that koalas are reservoirs of these opportunistic zoonotic organisms.

A range of bacterial pathogens have been reported to be involved in infectious disease processes in koalas, both captive and wild. Although confounded by multiple changes in taxonomy and nomenclature, it is now clear that Chlamydia pecorum and Chlamydia pneumoniae are major pathogens associated with a range of infectious diseases in both captive and wild koalas (Polkinghorne et al. 2013). Several other bacteria have been linked with infectious disease processes in koalas, including pneumonia associated with Bordetella bronchiseptica (McKenzie et al. 1979) and Actinomyces sp. (Stephenson et al. 2021); pleuritis associated with Streptobacillus moniliformis (Russell and Straube 1979); and ulcers caused by Mycobacterium ulcerans (Mitchell et al. 1987). Similar to other animals, koalas can suffer from a range of dermatological diseases as a result of bacterial, fungal, and parasitic infections (Blanshard and Bodley 2008). External ear infection (otitis externa) caused by Escherichia coli, Proteus spp., Pseudomonas aeruginosa, Pseudomonas fluorescens, and yeasts have been reported (Blanshard and Bodley 2008). However, our two cases of infections involving Lonepinellalike organisms appear to be the first recognition that these organisms can be involved in skin and soft tissue infection in the koala.

Veterinarians and diagnostic laboratories dealing with infected wounds of koalas (both captive and wild) need to be aware of the potential for the infections to be linked with Lonepinella-like organisms. There are a number of key flags that suggest the possibility of an organism being a member of the genus Lonepinella: a Gram-negative organism, oxidase positive (or weakly positive), catalase negative, and with a colony morphology typical of the family Pasteurellaceae. A failure of the common commercial identification systems to identify an organism with these phenotypic properties confidently is further support for the possibility of an isolate being a Lonepinella sp.

Genomic studies have suggested that bacteria apparently similar to Lonepinella koalarum may play a role in koala health and help protect against chlamydial infection (Dahlhausen et al. 2018). Although there are no formal standardized methods for determining antimicrobial resistance of Lonepinella-like organisms nor any interpretation guidelines, the data from Sinclair et al. (2019) suggest that these organisms are sensitive to a wide range of agents, with two of the four human bite isolates together with L. koalarum type strain showing possible resistance only to penicillin. Our search of the draft genome assemblies of the two new koala isolates, using the public databases ResFinder (Zankari et al. 2012) and CARD (Jia et al. 2017), did not detect any of the known resistance genes. More detailed analysis including the minimum inhibitory concentration of different antimicrobials should be done to determine the possibility of antimicrobial resistance of such organisms.

We acknowledge the Research Computing Centre (RCC) of the University of Queensland for supporting this research.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-2-00096.