Diversity and shifts of the bacterial community associated with Baikal sponge mass mortalities

Abstract

The disease of freshwater sponges was first discovered in 2011, when pink samples were found in the Central Basin of Lake Baikal. Subsequently, the visible signs of the disease have changed, and now sponges appear with various symptoms of damage to the body, such as discoloration, tissue necrosis, the formation of brown patches and dirty-purple biofilms on some branches. These signs of the disease are accompanied by the mass death of sponges. We identified differences in microbiomes by sequencing 16S rRNA genes and found changes in the consortium of microorganisms of freshwater Baikal sponges. We found that the observed imbalance in the studied microbial communities of diseased sponges is caused by several different conditionally pathogenic microorganisms that increase their negative effect by acting together and in concert, which leads to the death of photosynthetic microalgae and sponges. Sponges are an important component of coastal communities, and the massive loss of sponges can obviously affect the structure of benthic communities and the purity of water.

Fig 1. Dendrograms representing a degree of proximity between microbiomes compositions.

(A) Rank-based Kendall measure. (B) Presence-based unweighted UniFrac measure. Abundance values in the samples were used at the family level; dendrograms were constructed using UPGMA clustering.

Fig 2. Relative abundances of bacterial groups in the microbiomes of healthy and diseased sponges.

Bubble chart is shown at the class level for the 2010, 2011 and 2015 samples of sponges 2010, 2011 and 2015.

Fig 3. Heatmap of 25 most abundant bacterial families.

The relative abundances for individual 2015 samples are averaged. The abundance values were transformed by the quantile normalization, to reduce the biases introduced by sequencing technology. The right column shows a significance of variation between groups, estimated using ANOVA test; width of the bars corresponds to–ln (p-value). Red line separates the significance level of 0.999 (p-value < 0.001).

Fig 4. A comparative presentation of the 2015 samples.

Result given for (A) heatmap of the 15 most abundant bacterial groups, at the genus level. The right column on the right shows a significant difference between the healthy and diseased samples, estimated using Mann-Whitney test; width of the bars corresponds to–ln (p-value). Red line separates the significance level of 0.99 (p-value < 0.01). (B) Correspondence analysis at the OTU level based on the unweighted UniFrac measure as a distance.

Fig 5. An increase in heterogeneity of diseased samples.

Biodiversity values for the samples used in the study (A) Shannon distribution and (B) Simpson distribution. Bars indicate the values for a composition of aggregated communities.