Bacteria and Metabolic Potential in Karst Caves Revealed by Intensive Bacterial Cultivation and Genome Assembly

Abstract

Karst caves are widely distributed subsurface systems, and the microbiomes therein are proposed to be the driving force for cave evolution and biogeochemical cycling. In past years, culture-independent studies on the microbiomes of cave systems have been conducted, yet intensive microbial cultivation is still needed to validate the sequence-derived hypothesis and to disclose the microbial functions in cave ecosystems. In this study, the microbiomes of two karst caves in Guizhou Province in southwest China were examined. A total of 3,562 bacterial strains were cultivated from rock, water, and sediment samples, and 329 species (including 14 newly described species) of 102 genera were found. We created a cave bacterial genome collection of 218 bacterial genomes from a karst cave microbiome through the extraction of 204 database-derived genomes and de novo sequencing of 14 new bacterial genomes. The cultivated genome collection obtained in this study and the metagenome data from previous studies were used to investigate the bacterial metabolism and potential involvement in the carbon, nitrogen, and sulfur biogeochemical cycles in the cave ecosystem. New N₂-fixing Azospirillum and alkane-oxidizing Oleomonas species were documented in the karst cave microbiome. Two pcaIJ clusters of the β-ketoadipate pathway that were abundant in both the cultivated microbiomes and the metagenomic data were identified, and their representatives from the cultivated bacterial genomes were functionally demonstrated. This large-scale cultivation of a cave microbiome represents the most intensive collection of cave bacterial resources to date and provides valuable information and diverse microbial resources for future cave biogeochemical research.IMPORTANCE Karst caves are oligotrophic environments that are dark and humid and have a relatively stable annual temperature. The diversity of bacteria and their metabolisms are crucial for understanding the biogeochemical cycling in cave ecosystems. We integrated large-scale bacterial cultivation with metagenomic data mining to explore the compositions and metabolisms of the microbiomes in two karst cave systems. Our results reveal the presence of a highly diversified cave bacterial community, and 14 new bacterial species were described and their genomes sequenced. In this study, we obtained the most intensive collection of cultivated microbial resources from karst caves to date and predicted the various important routes for the biogeochemical cycling of elements in cave ecosystems.

Keywords: 3-oxoadipate-CoA transferases; Azospirillum; Oleomonas; bacterial cultivation; biogeochemical cycling; karst cave microbiome.

FIG 1

(a) Workflow of the isolation procedure and the diversity of the cultured cave bacteria. (b and c) Boxplots show the Shannon indices of the cultivated bacterial strains from the two caves and the three cave niches (rock, sediment, and water). The pie charts in panels b and c show the taxonomy distribution of the cave isolates from the two caves and the three cave niches. (c) PCoA plot shows the β-diversity of the cultured cave bacteria based on the Bray-Curtis dissimilarity; the Venn diagram shows the intersection of the cave isolates from the cave niches at the species level.

FIG 2

Taxonomic distribution of the cultured cave bacterial collection and its representativeness in 16S rRNA gene amplicon data sets. (a) Taxonomic distributions at the phylum and genus levels. Proteo, Proteobacteria; Actino, Actinobacteria; Firmi, Firmicutes. (b) Boxplots show the percentages of the sequences in the amplicon data sets that are represented by the cultured isolates, and the triangles in each boxplot indicate the mean representativeness of each data set.

FIG 3

Morphologies and phylogenetic affiliations of the new species isolated from the cave samples. (a) Morphology from transmission electron microscopy. (b) The phylogenetic tree was constructed based on the 16S rRNA genes using the neighbor-joining algorithm.

FIG 4

Metabolic overview of the newly isolated bacterial species from the caves. (a) Assimilation of the carbon sources according to the Biolog GEN III system; purple indicates positive and white indicates negative. (b) Distributions of the COGs in the 14 newly sequenced genomes; the COGs are color coded, with the highest number of genes shown in pink and the genes with the lowest number shown in green.

FIG 5

Overview of the metabolisms of the cave cultured genome collection (a, c, and e) and the public cave metagenome data (b, d, f) and their relationships to the C/N/S cycles. The numbers and percentages on the arrows in panels a, c, and e represent the number of species that are able to perform the conversion and their relative abundances; the width of the arrow is in proportion to the number of species that are able to perform the transformation. The color ranges in panels b, d, and f indicate the transcripts per million (TPM) values of each KO in the metagenome data (accession numbers are shown as x axis labels).

FIG 6

Representative genetic clusters (a) and β-ketoadipate pathway (b), and the two 3-oxoadipate-CoA-transferase gene clusters (c) and their enzymatic activity in the pathway (d). The red percentages in panel a indicate the amino acid similarities between the cave isolates and strain ATCC 35469. The controls in panel d summarize three conditions: the assay mixture without enzyme, with K5PcaIJ but without succinyl-CoA, or with 30PcaIJ but without succinyl-CoA.