Diversity and transmission of koala retrovirus: a case study in three captive koala populations

KoRV genetic diversity

Illumina env amplicon deep sequencing was carried out on genomic DNA isolated from the peripheral blood mononuclear cells (PBMCs) of 64 koalas housed in three captive koala colonies within Australia, one in South-East Queensland (SE QLD; colony C, n = 33) and two in Sydney, New South Wales (NSW; colony D, n = 14; colony E, n = 17). The hyper-variable receptor binding domain within the env gene was amplified which, following quality control and clustering at 97% identity, resulted in the detection of 93 unique KoRV sequences (deposited to GenBank; Table 1). Proviral genomic DNA isolated from PBMCs was used as this provides the highest detection rate for the exogenous subtypes infecting a koala compared to other sample types3,11.

Table 1 KoRV subtype diversity in captive koala populations.

The read count was found to vary drastically between individuals with 19,136 KoRV reads present on average, ranging from 2280 to 64,243, after filtering and quality control. Protein alignment of the 93 in silico translated sequences to those of known KoRV subtypes led to their classification as one of eight subtypes (A, B, D, E, G, H, I and K). Consistent with other studies, KoRV-A was found to be the most prevalent subtype detected, identified in 100% of the koalas analyzed from all three populations (Table 1). This was followed closely by KoRV-D which was detected in 53 (82.8%) out of 64 koalas examined. Unsurprisingly, these two subtypes also had the greatest number of unique sequences detected with 25 KoRV-A and 30 KoRV-D sequences identified across all three institutions. KoRV-G and H were found to be the least prevalent subtypes, detected in one and two koalas, respectively. KoRV-E, G and K were found to have the least genetic diversity, with only two sequences detected for each across all KoRV-E (n = 4), G (n = 1) and K (n = 15) positive koalas (Table 1). KoRV-C, F, L and M were not identified within these populations.

Whilst a small proportion of sequences were only identified within a single koala, highlighting the continual within-host evolution of this virus, the majority (84.9%) of sequences were detected in two or more koalas (Table 1; Supplementary Table S2). Consistent with previous research, the full-length endogenous KoRV-A sequence (GenBank accession number AF151794) was identified within all analyzed koalas where it accounted for 98.2% of an individual’s reads on average, ranging between 80.4% and 100%. The next most within koala abundant sequence was a KoRV-B sequence (GenBank accession number ON839123) detected in 16 (25%) individuals across all three populations. This sequence accounted for 0.27% of a koala’s total KoRV reads on average, ranging between 0% and 7.4%, and 23.5% of a koala’s total KoRV-B reads on average, ranging between 0% and 99.8% (Supplementary Table S2). KoRV subtypes I and K were each found to have dominant sequences that were detected within 81% of KoRV-I (ON839175) and 100% of KoRV-K (ON839182) positive koalas, accounting on average for 41% and 97% of KoRV-I and K total reads, respectively. No dominant sequence was identified within KoRV-D positive koalas with the most prevalent sequence (ON839157) detected in less than 50% of positive koalas.

KoRV-A was the most abundant subtype in all analyzed koalas, representing 98.3% of KoRV reads on average, ranging from 82.3% to 100%. The other subtypes were comparatively reduced, accounting for less than 20% of reads in total (Fig. 1). The number of reads attributable to each KoRV subtype for each koala is shown in Fig. 1 and Supplementary Table S1. KoRV A, B and D were the only subtypes detected in all three populations, with KoRV-A and D identified at a similar prevalence and abundance across all three. Whilst KoRV-B was found to be more prevalent within colony E (64.7% of koalas), it was most abundant in colony D where it accounted for 4% of an animal’s total KoRV reads on average (Fig. 1B). KoRV-I was found to be more prevalent in colony E (64.7% of koalas) in comparison to colony C (30.3% of koalas), however, despite an obvious outlier, it was found at a similar abundance across both populations. The other exogenous subtypes were detected variably among the populations (Fig. 1B). Overall, colony C was found to have the least genetic diversity with KoRV-A accounting for 99% of total KoRV reads on average (Fig. 1A).

Figure 1
figure 1

Percentage of KoRV reads grouped by subtypes. Prevalence of KoRV subtypes in genomic DNA from koalas housed at colony C (n = 33), colony D (n = 14) and colony E (n = 17). (A) Subtype abundance for each animal is shown for all populations. Colours indicate the different subtypes detected. (B) Percentage relative abundance for each subtype is summarized for the three populations. Each point represents an individual koala with the mean ± SD shown. N.D not detected.

At least two subtypes were detected in 55 (85.9%) of the 64 koalas analyzed, with a maximum of four subtypes identified in 23 koalas. Notably, the remaining nine koalas only had KoRV-A (Fig. 1A; Supplementary Table S1; C2, C14, D8, D10, D11, D13, D14, E10 and E14). Whilst these individuals were housed in each of the three institutions, more than 50% resided in colony D in NSW. Interestingly, this colony was found to have the least number of subtypes detected of all three populations and at a relatively low abundance (Fig. 1B).

Two neonate tissue samples from colony D were also analyzed for their KoRV genomic DNA composition. Both samples were collected opportunistically after neonates died naturally failing to make it into the pouch following birth. The endogenous KoRV-A sequence (AF151794) was found in both neonate 1 (D15) and 2 (D16), where it accounted for 99.98% and 100% of their total KoRV reads, respectively. The remaining 0.02% of KoRV reads for neonate 1 were attributable to a KoRV-D sequence. This sequence was also detected at low levels in the dam (D1), sire (D17) and both paternal grandparents (D9 and D12) of this neonate (Supplementary Fig. S1), however, not in the maternal grandmother (D13) sample. Notably, these animals were analyzed via blood (D1, D9, D12–13) and spleen tissue (D17). Both neonates shared the same dam, however the maternal grandfather and neonate 2 sire were not included in this study.

KoRV exogenous transmission

The number of KoRV sequences shared between unrelated (n = 426), maternally related (m-related; n = 100), paternally related (p-related; n = 156), dam-joey (n = 40), sire-joey (n = 16) and mating partner (n = 17) koala pairs was compared between pair types by fitting generalised linear mixed models using the MCMCglmm 2.29 package in R35. A maternal lineage was defined as those related through a strictly female line. Due to the overall low sequence diversity (and consequently low number of sequences shared between koala pairs) in these populations, all KoRV subtypes were combined for analysis to provide adequate data for model fitting. This analysis did not include subtypes E, G or H, as they were detected in too few animals for meaningful comparisons to be made. The endogenous KoRV-A sequence was also omitted, as it was shared among all animals.

Overall, dam-joey pairs were found to share significantly more sequences than unrelated and sire-joey pairs, sharing 1.9 sequences on average (95% credible interval (CI) 1.5–2.4) compared to 1.3 (95% CI 0.9–1.9) for unrelated (Fig. 2; p < 0.01) and 1.2 (95% CI 0.7–1.8) for sire-joey (p < 0.05). Similar findings were observed when comparing m-related pairs (averaged 1.9 sequences shared; 95% CI 1.6–2.3) with unrelated pairs (Fig. 2; p < 0.001) and p-related pairs (averaged 1.6 sequences shared; 95% CI 1.3–1.8; p < 0.05). Whilst significantly more sequences were shared by p-related koalas than unrelated koalas, this was not as notable as dam-joey or m-related pairs (Fig. 2; p < 0.05). No evidence of sexual transmission was observed, with mating partners (averaged 0.8 sequences shared; 95% CI 0.5–1.3) found to share significantly less sequences on average compared to unrelated koala pairs (Fig. 2; p < 0.05).

Figure 2
figure 2

Average sequences shared with 95% credible intervals from generalised linear mixed models of subtype sharing. The expected number of shared KoRV-A, B, D, I and K sequences is shown for paternally related (P-related; n = 156), maternally related (M-related; n = 100), sire-joey (n = 16), dam-joey (n = 40) and mating partner (n = 17) koala pairs. Expected sharing between unrelated (n = 426) koalas is represented by a dashed line with 95% credible intervals highlighted in grey. Maternal relatives were defined by those related through a strictly female lineage. The original, endogenous KoRV-A sequence was omitted from this analysis. Asterisks above nodes indicate significance to the unrelated reference group. Significance between different pair groupings is shown on the right. *p < 0.05, **p < 0.01, ***p < 0.001.

Whilst not included in the statistical modelling due to low sample size (n = 4), evidence of dam to joey transmission of KoRV-E was also found within colony D. In this population, KoRV-E was only identified within the offspring (D5 and D6) of two unrelated, KoRV-E positive dams (D3 and D7, respectively). Notably, however, there were no matched samples from KoRV-E positive sires in these populations.

Read more here: Source link