Characterisation of the koala (Phascolarctos cinereus) pouch microbiota in a captive population reveals a dysbiotic compositional profile associated with neonatal mortality

Abstract

Background: Captive koala breeding programmes are essential for long-term species management. However, breeding efficacy is frequently impacted by high neonatal mortality rates in otherwise healthy females. Loss of pouch young typically occurs during early lactation without prior complications during parturition and is often attributed to bacterial infection. While these infections are thought to originate from the maternal pouch, little is known about the microbial composition of koala pouches. As such, we characterised the koala pouch microbiome across the reproductive cycle and identified bacteria associated with mortality in a cohort of 39 captive animals housed at two facilities.

Results: Using 16S rRNA gene amplicon sequencing, we observed significant changes in pouch bacterial composition and diversity between reproductive time points, with the lowest diversity observed following parturition (Shannon entropy - 2.46). Of the 39 koalas initially sampled, 17 were successfully bred, after which seven animals lost pouch young (overall mortality rate - 41.18%). Compared to successful breeder pouches, which were largely dominated by Muribaculaceae (phylum - Bacteroidetes), unsuccessful breeder pouches exhibited persistent Enterobacteriaceae (phylum - Proteobacteria) dominance from early lactation until mortality occurred. We identified two species, Pluralibacter gergoviae and Klebsiella pneumoniae, which were associated with poor reproductive outcomes. In vitro antibiotic susceptibility testing identified resistance in both isolates to several antibiotics commonly used in koalas, with the former being multidrug resistant.

Conclusions: This study represents the first cultivation-independent characterisation of the koala pouch microbiota, and the first such investigation in marsupials associated with reproductive outcomes. Overall, our findings provide evidence that overgrowth of pathogenic organisms in the pouch during early development is associated with neonatal mortality in captive koalas. Our identification of previously unreported, multidrug resistant P. gergoviae strains linked to mortality also underscores the need for improved screening and monitoring procedures aimed at minimising neonatal mortality in future. Video Abstract.

Keywords: Dysbiosis; Endangered species; Enterobacteriaceae; Klebsiella pneumoniae; Koala; Marsupial; Muribaculaceae; Phascolarctos cinereus; Pluralibacter gergoviae; Pouch; Reproduction.

Fig. 1

Timeline displaying sample collection relative to neonatal developmental phases and reproductive cycle. Entire timeline is representative of 12 months, with major developmental events displayed above marked by red arrows and sampling time points marked with circles below. Opportunist time point samples, noted in orange and with an asterisk, were not collected in all longitudinally sampled dams

Fig. 2

Swab collection from koala pouches. A Prior to sample collection, koalas are gently restrained by trained personnel using the fore and hindlimbs to prevent injury. B To collect samples, the pouch is propped open using hair on the anterior edge of the pouch, and the swab is carefully inserted into the pouch perpendicular to the unoccupied teat. Samples were collected from the area displayed in (C)

Fig. 3

Bacterial community diversity in the pouch microbiota of captive koalas. Alpha- and beta-diversity metrics were calculated on a feature table rarefied to 1000 reads per sample. Alpha-diversity metrics included A Shannon diversity and B observed species (or ASVs), with Kruskal–Wallis testing performed to identify significant differences between groups. Beta diversity is presented as weighted UniFrac distances, with ordination performed via principal coordinates analysis (PCoA). C Displays ordination on a 2-dimensional PCoA plot with PC1 and PC2, while D displays data plotted to PC1 and PC3. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.00001

Fig. 4

Phylum- and family-level bacterial composition in the pouch microbiota of captive koalas. The figure displays A phylum-level and B family-level bacterial composition in individual pouch samples grouped by reproductive time point. Values on both plots represent relative abundances of taxa. A Visualised using QIIME2 (v2021.4). B Visualised using ggplot2. VEL, very-early lactation; EL, early lactation; PE, post emergence

Fig. 5

Bacterial community beta diversity stratified by breeding status. Ordination and principal coordinates analysis (PCoA) was performed using weighted UniFrac distances calculated on data rarefied to 1000 reads per sample. A Displays PCoA ordination of pouch samples faceted by time point at anoestrus, very-early lactation, and early lactation. B Displays ordination of early lactation, loss, and postemergence samples only using PC1 and PC2 axes

Fig. 6

Phylum-level composition of the pouch microbiome at each reproductive time point grouped by breeding status. Samples were split by breeding group, where BG1 = successful breeders and BG2 = unsuccessful breeders. Plot was produced using QIIME2 (v2021.4)

Fig. 7

Family-level differences and differentially abundant genera in the pouch microbiota of successful vs. unsuccessful breeding koalas during early lactation. A Displays relative abundance bar plot of EL samples grouped by breeding status at the family level. B Differentially abundant ASVs (FDR-adj p < 0.05) in successful and unsuccessful breeder EL samples were identified using pairwise analysis in DeSeq2 (v3.14). ASVs were assigned taxonomy at the genus (y-axis) and family (as per bar plot legend) level, with each point representing individual ASVs within genera. The x-axis shows the ‘log2 fold change’ in taxa between groups, with negative values representing higher abundance in successful breeders and positive values representing higher abundance in unsuccessful breeders. Visualisation was performed using QIIME2 (v2021.4) and ggplot2 (v3.3.5). Raw data presented in Supplementary Tables 5, 6, 7, 8, 9, 10 and 11

Fig. 8

Genus-level bacterial community composition in pouch samples collected following loss of pouch young. The figure displays the relative abundance of bacterial genera in individual loss samples. Visualisation was performed using QIIME2 (v2021.4)