Page 16 - Read Online

Page 10 of 16 Li et al. Microbiome Res Rep 2024;3:26 https://dx.doi.org/10.20517/mrr.2023.57

For labs without automation capacity, 96-well plate-based manual digestion and desalting procedures can
also be used [Supplementary Method 3]. However, we validated that the automation increased
reproducibility and sensitivity compared to manual liquid handling [Supplementary Figure 2]. Samples are
TM
then labeled using TMT11plex reagents pre-aliquoted and dried into 96-well plates [Figure 1E]. TMT
labeling is quenched, pooled by rows, and desalted prior to sample analysis using an LC-MS/MS. *.RAW
files are subjected to database search using MetaLab software, and subsequent data analysis is preformed
[Figure 1F]. A timeline breakdown for the whole protocol is given in Figure 1G.

Proof-of-concept of the RapidAIM 2.0 approach
Here, we exemplify the described RapidAIM 2.0 workflow with a demonstration study, in which we
evaluated the use of different human stool sample collection and biobanking processing workflows on
individual gut microbiome responses to kestose, a prebiotic oligosaccharide as of established knowledge of
[34]
its impact on gut metaproteomes [Figure 2A]. All processed fecal samples were stored at -80 °C before
performing the RapidAIM 2.0 experiment. Samples of all comparisons were randomized before being
TM
labeled using TMT11plex . An Ultimate 3000 RSLCnano system coupled to an Orbitrap Exploris 480 was
used for the analysis of TMT-labeled samples using a two-hour gradient. Details of LC-MS/MS parameters
are as described in Equipment setup. LC-MS/MS *.RAW files were searched against the IGC database [37]
using MetaLab 2.3.0.

Altogether, 162 samples were labeled with a TMT11plex kit, resulting in 17 multiplexed samples. On
TM
average, 17,078 ± 24 MS/MS spectra were identified (MS/MS identification rate 24.2% ± 1.0%), resulting in
14,026 ± 570 identified peptides and 5,014 ± 142 protein groups per multiplexed sample set (Mean ± SD,
N = 17; Figure 2B-D; Supplementary Figure 3). This is highly comparable to the previous label-free
[18]
quantification in RapidAIM . A substantial proportion (97.3%) of identified peptides belong to bacteria,
with the other 2.7% assigned to Eukaryote, specifically human proteins [Supplementary Figure 4]. Dataset
was then preprocessed using MSstatsTMT which includes a pipeline of spectrum-level normalization,
protein summarization, and protein-level normalization using the LogSum method . Multivariate
[38]
statistical analysis can then be performed using the processed dataset.

We first compared four different sample processing methods on microbiome responses (see Methods).
Among whese, the Gauze method has the merit of low cost, the Vacuum filtration method has the merit of
rapid processing of large volumes, the Spin method has the merit of easy handling, and the 100 g method
has the merit of both low cost and easy handling. Principal component analyses (PCA) show that all
technical triplicates were well-clustered [Figure 2E]. PCA and hierarchical clustering [Figure 2E and F] also
show that for all three tested individual microbiomes, samples were clustered by individual and condition (0
day baseline, blank, and kestose). No separation between the four different stool sample processing
strategies was observed, suggesting that all four strategies were applicable for RapidAIM without impact on
microbiome functional responses and that one strategy can be selected based on the evaluation of cost and
time preferences.

We next assessed the feasibility of preserving samples in specific collection buffers and storing them before
undergoing biobanking processing. We tested the storage efficacies of our in-house PBS-glycerol buffer and
the commercial GutAlive buffer by comparing samples processed immediately after collection (0 day) with
those stored for 24 and 72 h. While the sample collected using the in-house buffer were stored at 4 °C, the
GutAlive samples were stored at room temperature [Figure 3A]. Similar to the previous test on sample
processing method, the sample storage test showed clear separations by storage conditions. While samples
stored for 24 h under room temperature in the GutAlive kit showed no separation from the samples that