Inferring replication states of bacteria and viruses in enrichment cultures via long-read sequencing

Abstract

Most microorganisms cannot be cultured in isolation, necessitating sophisticated methods for studying their (eco)physiology. While numerous approaches can probe the activity of given microbes in enrichment cultures, no single technique can render simultaneous data on both metabolic capacities and mobile genetic elements. Here, we apply long-read sequencing to monitor the incorporation of non-canonical bases in genome-resolved metagenomic datasets and elucidate the replication patterns of both bacteria and phages. This technology enables the simultaneous reconstruction of both prokaryotic and viral genomes (alongside genomics downstream analyses like metabolic predictions), in addition to providing information regarding their replication in enrichment cultures. By spiking the base analog 5-bromo-2'-deoxyuridine (BrdU) into activated sludge microcosms, we determined that 114 of the 118 high-quality genomes recovered were actively replicating in enrichment cultures from activated sludge and identified both slow (low BrdU incorporation and change in abundance) and rapidly replicating organisms (high BrdU incorporation and change in abundance). Some of the genomes detected exhibited regions rich in BrdU that were predicted to represent prophages in their lytic cycle. Ultimately, this novel means of monitoring the replication responses of microbes, and deciphering their genomes and active mobile genetic elements will advance and empower strategies aimed at isolating previously uncultivated microbes in pure culture.

Keywords: 5-bromo-2′-deoxyuridine (BrdU); enrichment cultures; genome-resolved metagenomics; nanopore sequencing; prokaryotic replication; prophage induction.

Figure 1

Experimental design and BrdU incorporation patterns of 20 most abundant organisms. (A) Experimental design of incubations with and without BrdU and respective data analysis workflow. Activated sludge was incubated with and without BrdU for a total of 56 h, during which samples were collected at seven distinct time points. Illumina short-reads and Nanopore long-reads were generated and hybrid-assembled into dereplicated MAGs, whose differential coverage was calculated across all samples. BrdU substitutions were called on long-reads using DNAscent [9] and used to calculate scaffold- and genome-level BrdU containment. (B) GTDB phylogeny (tree, colored branch labels) and taxonomic affiliation at phylum- and class-level of the 20 most abundant MAGs according to the median values for genome-level normalized BrdU accumulation across replicates. Labels are provided for each MAG given their lowest known taxonomic rank (y-axis of bubble plot). Median normalized (see Supplementary Methods—BrdU count normalization) relative BrdU counts across replicates (bubble color) and coverage (bubble size) of MAGs across time points (x-axis) are displayed in a bubble plot.

Figure 2

Examples of distribution of BrdU incorporation across bacterial and prophage genomes. Genome-wide (x-axis, split into 250 histogram aggregates of ~36 kbp each) BrdU accumulation (bar lengths, y-axis) across time points (bar colors), with vertical dashed line denoting the predicted origin of replication. The displayed MAG is AcBaMe_P2T5_Deltaproteobacteria_bacterium_69_7. (A). Visualization of BrdU hits (open circles) across the two prophage- carrying scaffolds (y-axis) found to have the highest differences in BrdU incorporation (filled circles in genomic scaffolds) between the genomic and prophage portion of the scaffold (pink lines below the density curve, B). Annotations found for each prophage are expanded across the prophage genomes (x-axis indicates genomic position). Colors and labels indicate annotation, while in the background in gray, a density curve plots the density of BrdU hits (asterisks) across the prophage genomes.