Evaluating live microbiota biobanking using an ex vivo microbiome assay and metaproteomics

Abstract

Biobanking of live microbiota is becoming indispensable for mechanistic and clinical investigations of drug-microbiome interactions and fecal microbiota transplantation. However, there is a lack of methods to rapidly and systematically evaluate whether the biobanked microbiota maintains their cultivability and functional activity. In this study, we use a rapid ex vivo microbiome assay and metaproteomics to evaluate the cultivability and the functional responses of biobanked microbiota to treatment with a prebiotic (fructo-oligosaccharide, FOS). Our results indicate that the microbiota cultivability and their functional responses to FOS treatment were well maintained by freezing in a deoxygenated glycerol buffer at -80°C for 12 months. We also demonstrate that the fecal microbiota is functionally stable for 48 hours on ice in a deoxygenated glycerol buffer, allowing off-site fecal sample collection and shipping to laboratory for live microbiota biobanking. This study provides a method for rapid evaluation of the cultivability of biobanked live microbiota. Our results show minimal detrimental influences of long-term freezing in deoxygenated glycerol buffer on the cultivability of fecal microbiota.

Keywords: Biobanking; freezing; gut microbiota; metaproteomics; microbiome assay.

Figure 1.

Evaluation of the cultivability of frozen biobanked fecal microbiota. (a) experimental workflow. Fresh stool was directly collected from healthy adult volunteers into pre-reduced 10% glycerol in phosphate buffered saline (PBS). Samples were transferred into anaerobic workstation immediately, homogenized to make a 20% (w/v) fecal slurry and aliquoted followed by filtering using sterile gauzes to remove large particles. The pre-processed stools were cultured directly in anaerobic workstation or frozen in −80°C for up to 52 weeks prior to culturing with or without fructo-oligosaccharide (FOS) in triplicates. (b) Principal component analysis (PCA) score plot of quantified protein groups. The log10-tranformed LFQ intensities of protein groups that were quantified in >50% samples were used for PCA analysis. Different symbol colors indicate the weeks of frozen in −80°C prior to culturing, and different shapes indicate the treatment groups for each culturing experiment. Baseline, uncultured microbiome samples; QC, quality controls for monitoring mass spectrometry measurement performance; V31, V33 and V34, origin donors of the microbiome samples.

Figure 2.

Principal component analysis (PCA) score plot of quantified functional orthologs of baseline and cultured microbiomes using fresh or frozen biobanked stools. The log10-tranformed LFQ intensities of COGs that were quantified in >50% samples were used for PCA analysis. Different symbol colors indicate the weeks of frozen in −80°C prior to culturing, and different shapes indicate the treatment groups for each culturing experiment. Baseline, uncultured microbiome samples; FOS, fructo-oligosaccharide treatment group; PBS, phosphate buffered saline treatment group. QC, quality controls for monitoring mass spectrometry measurement performance; V31, V33 and V34, origin donors of the microbiome samples.

Figure 3.

Abundance distribution of quantified COG categories in baseline and cultured microbiomes using fresh or frozen biobanked stools. Group average of LFQ intensities were used for plotting. Baseline, uncultured microbiome samples; FOS, fructo-oligosaccharide treatment group; PBS, phosphate buffered saline treatment group. V31, V33 and V34, origin donors of the microbiome samples. Different colors indicate COG categories. Column facet grid names (wk0, 1, 2, 4, 8, 16, 24 and 52) indicate the weeks of frozen in −80°C prior to culturing.

Figure 4.

Phylum level composition of baseline and cultured microbiomes using fresh or frozen biobanked stools. Group average of relative abundances was used for plotting. Baseline, uncultured microbiome samples; FOS, fructo-oligosaccharide treatment group; PBS, phosphate buffered saline treatment group. V31, V33 and V34, origin donors of the microbiome samples. Different colors indicate phyla as indicated in the legend. Column facet grid names (wk0, 1, 2, 4, 8, 16, 24 and 52) indicate the weeks of frozen in −80°C prior to culturing.

Figure 5.

Relative abundances of Actinobacteria in baseline and cultured microbiomes using fresh or frozen biobanked stools. Baseline, uncultured microbiome samples; FOS, fructo-oligosaccharide treatment group; PBS, phosphate buffered saline treatment group. V31, V33 and V34, origin donors of the microbiome samples. Column facet grid names (wk0, 1, 2, 4, 8, 16, 24 and 52) indicate the weeks of frozen in −80°C prior to culturing. Box plots were generated using ggplot2 with data points overlapped with the boxes. The first and third quantiles were indicated as the box width and median was also displayed in the middle of box. Upper and lower whiskers indicate the smallest value ≥1.5 interquartile range (IQR) and largest value ≤1.5 IQR, respectively.

Figure 6.

Effects of delayed fecal sample processing on the cultivability of microbiomes. (a) Principal component analysis (PCA) score plots of quantified protein groups. Baseline, uncultured microbiome samples; FOS, fructo-oligosaccharide treatment group; PBS, phosphate buffered saline treatment group. V31, V33 and V34, origin donors of the microbiome samples; Fresh, culturing with fresh stools; ICE6h, culturing with stools stored on ice for 6 hours; DRYICE6h, culturing with stools stored on dry ice for 6 hours. (b) Phylum level taxonomic compositions as calculated using metaproteomic data. Group average of relative abundances was used for plotting. NA indicated no treatment/culturing for baseline samples. Different colors indicate different phyla as shown in the legend at the bottom.

Figure 7.

Influence of two-day delayed processing on the cultivability of microbiomes. (a) Principal component analysis (PCA) score plots of quantified protein groups. FOS, fructo-oligosaccharide treatment group; PBS, phosphate buffered saline treatment group. V51, origin donor of the microbiome samples; Duration of 0 hr, 24 hr and 48 hr (different colors) indicate the stool processing were processed immediately, delayed for 24 hours and 48 hours, respectively. (b) Phylum level taxonomic compositions as calculated using metaproteomic data. Group average of relative abundances was used for plotting. Different colors indicate different phyla as shown in the legend at the bottom.