Significant progress in microbiome characterization has revealed that microbial communities—whether buried under the ocean floor or nestled within our intestines—are even more biologically complex and consequential than expected. Human health studies are continually uncovering new associations between our microbiome and disease, such as recent findings that microbial populations influence the fate of certain cancers or that adjusting the makeup of a patient’s gut microbiome can reduce the severity of autism symptoms.
Next-generation sequencing tools have been used widely to catalog what strains are present in a given human gut microbiome (HGM) sample and in what proportions. However, unraveling these complex populations will require determining what each constituent strain does and how these strains interact with each other and their environment, because these interactions have important biological ramifications. Elucidating these functional properties requires cost-effective, high-throughput tools for microbial isolation and cultivation that can facilitate analysis of thousands of strains at once while simultaneously preserving them in living form for downstream functional studies.
Today’s microbiology workflow leans heavily on labor-intensive, low-throughput tools such as Petri dishes, agar shake tubes, and broth cultures. These tools, which were developed in the 19th century, were never intended to enable the analysis of complex microbial communities with functions beyond those of single (noninteracting) members. New microbiology tools that allow researchers to efficiently investigate the functions of interacting microbial strains are required.
High-throughput isolation, cultivation, and screening
The Prospector™ system is designed to perform cultivation at the scales needed to characterize entire populations of microbes from the HGM. The system can be used to cultivate hundreds to thousands of bacteria from HGM samples in just 1 week with an automated workflow that significantly reduces hands-on time compared to traditional methods.
The core of the technology is a highly dense array containing over 6000 nanoscale microwells, each of which represents a 3-nL cultivation chamber. Loading an array with an appropriately diluted HGM sample results in microwells containing a single bacterium, which means that each strain can be cultured in isolation from other strains.
Clonal cultures can then be grown across thousands of microwells simultaneously, providing ready access to a large and diverse library of isolates, including rare isolates that cannot be outcompeted in the array microwell, at a throughput not possible with classical workflows. In addition, given the ease and speed with which multiple parallel experiments can be conducted with Prospector, it is practical to increase library diversity by strategically testing several different media.
The instrument’s integrated optics enable fluorescent or colorimetric detection of array microwells with living isolates via growth indicators. Fluorescence images of the array with red (635 nm), green (523 nm), and blue (488 nm) excitations can be captured. The instrument then picks and transfers target isolates to a standard 96- or 384-multiwell plate in a fully automated manner.
Cultivation on the Prospector can be done with an easy three-step workflow that starts with the preparation of a microbial suspension, loading and incubation of the array, and automated imaging and recovery of target isolates (Figure 1). The Prospector instrument readily fits into an anaerobic chamber (Coy Lab Products and Anaerobe Systems) such that all steps can be done without samples ever leaving the chamber and without operators needing to manage dozens or hundreds of Petri dishes. The Prospector workflow to culture HGM samples uses resorufin as an indicator of anaerobic metabolism. The biological reduction of resorufin to dihydroresorufin by metabolic byproducts of anaerobic fermentation occurs more rapidly than does the abiotic reduction of resorufin by molecular hydrogen, allowing empty nanowells to be discriminated from those containing cultures.
This discrimination under anaerobic conditions is accomplished by looking at the change in signal for each nanowell from time zero to any later timepoint. Nanowells showing a greater decrease in fluorescence than the abiotic background signal change shown by empty wells are designated culture positive (Figure 2).
Six unique human fecal samples from healthy donors for which paired metagenomic data was available were cultured on the Prospector system, and 2790 isolates were obtained. Sanger sequencing of isolates indicated 5 phyla, 9 classes, 12 orders, 19 families, 27 genera, and 45 species (23 species for which at least six isolates were recovered (Figure 3). An additional 22 species were identified in the six samples for which there were five or fewer isolates (Table 1). Not only were these isolates relatively rare in terms of their capture in culture, but in nearly every case, paired metagenomics data had predicted a prevalence of ≤0.2% for these species or genera to be found in the samples.
Comparison of Sanger 16S rRNA sequencing data for identification at the species level with predicted prevalence from metagenomic data highlights the strength of Prospector in capturing rare members of a microbial community. Thirty-three species isolated from one sample represented 5 phyla, 8 classes, 10 orders, 16 families, and 20 genera (Table 2). Of these 33 species, 18 were predicted to be present at <1.0% abundance, demonstrating the ability of the system to isolate rare strains. Prospector was also able to isolate seven species not identified in the same sample by metagenomic analysis.
The ability to perform massively parallel cultivation on the Prospector can enable rapid generation of large, diverse libraries of living isolates that can be used for downstream functional studies. Data matched to metagenomic results demonstrates the use of the Prospector in culturing rare strains that may play a crucial role in the biology of the microbiome. Prospector’s miniaturization together with the automated workflow offers significant reductions in hands-on processing time, enabling scientists to gain back valuable time to do research.
With this previously unimaginable throughput, researchers can now ask questions about microbiomes that were never feasible before—and ultimately get results that will make it possible to harness the untapped potential of these complex microbial communities.
Jennifer Liang, Abelardo Arellano, and Sanam Sajjadi are research associates and Talia Jewell, PhD (firstname.lastname@example.org), is group lead, microbiology, at General Automation Lab Technologies (GALT).