In silico evaluation of a targeted metaproteomics strategy for broad screening of cellulolytic enzyme capacities in anaerobic microbiome bioreactors

Abstract

Background: Microbial-driven solubilization of lignocellulosic material is a natural mechanism that is exploited in anaerobic digesters (ADs) to produce biogas and other valuable bioproducts. Glycoside hydrolases (GHs) are the main enzymes that bacterial and archaeal populations use to break down complex polysaccharides in these reactors. Methodologies for rapidly screening the physical presence and types of GHs can provide information about their functional activities as well as the taxonomical diversity within AD systems but are largely unavailable. Targeted proteomic methods could potentially be used to provide snapshots of the GHs expressed by microbial consortia in ADs, giving valuable insights into the functional lignocellulolytic degradation diversity of a community. Such observations would be essential to evaluate the hydrolytic performance of a reactor or potential issues with it.

Results: As a proof of concept, we performed an in silico selection and evaluation of groups of tryptic peptides from five important GH families derived from a dataset of 1401 metagenome-assembled genomes (MAGs) in anaerobic digesters. Following empirical rules of peptide-based targeted proteomics, we selected groups of shared peptides among proteins within a GH family while at the same time being unique compared to all other background proteins. In particular, we were able to identify a tractable unique set of peptides that were sufficient to monitor the range of GH families. While a few thousand peptides would be needed for comprehensive characterization of the main GH families, we found that at least 50% of the proteins in these families (such as the key families) could be tracked with only 200 peptides. The unique peptides selected for groups of GHs were found to be sufficient for distinguishing enzyme specificity or microbial taxonomy. These in silico results demonstrate the presence of specific unique GH peptides even in a highly diverse and complex microbiome and reveal the potential for development of targeted metaproteomic approaches in ADs or lignocellulolytic microbiomes. Such an approach could be valuable for estimating molecular-level enzymatic capabilities and responses of microbial communities to different substrates or conditions, which is a critical need in either building or utilizing constructed communities or defined cultures for bio-production.

Conclusions: This in silico study demonstrates the peptide selection strategy for quantifying relevant groups of GH proteins in a complex anaerobic microbiome and encourages the development of targeted metaproteomic approaches in fermenters. The results revealed that targeted metaproteomics could be a feasible approach for the screening of cellulolytic enzyme capacities for a range of anaerobic microbiome fermenters and thus could assist in bioreactor evaluation and optimization.

Keywords: Anaerobic digester; Biogas; Glycoside hydrolases; Lignocellulose; Microbial community; Microbiome; Peptides; Targeted metaproteomics.

Fig. 1

Selection of unique peptides for GH families in the development of a targeted proteomics assay. A In this paper, the in silico selection of shared peptides (in rectangles) within proteins from specific GH families that are otherwise unique to entire groups of them, was demonstrated. Potential applications are: B monitoring the stable hydrolytic capabilities of an anerobic digester or condition-dependent changes over time or C evaluation of the hydrolytic potential of a bacterial community extracted from the environment for use as the starting inoculum for a digester according to the expression of specific families of GHs

Fig. 2

CAZymes annotated in the proteomes of different phyla. Box plots (left) show the percentage of CAZymes annotated in the proteomes of different bacterial and archaeal phyla using dbCAN2. The number of annotated MAGs per phylum are shown in parenthesis. Pie charts (right) show the relative fraction of different CAZyme classes, which include AAs (enzymes of the auxiliary activities), CBMs (carbohydrate-binding modules), CEs (carbohydrate esterases), GHs (glycoside hydrolases), GTs (glycosyltransferases), and PLs (polysaccharide lyases). Some proteins were also annotated with cohesin and S-layer homology domains, which are involved in the structure and formation of cellulosomes. MAGs lacking annotations at the phylum level are not shown

Fig. 3

Minimum number of unique tryptic peptides and their associated number of protein seeds in targeted GH families. A Total number of unique peptides selected for each GH family after comparison against other proteins in the biogas microbiome dataset and the number of protein seeds in which they are found. B Top 10 tryptic peptides (from blue bars in A) ranked by the highest number of protein seeds they cover in each GH family. C Percentages of total proteins in GH families covered by top 10, 50, 100, and 200 peptides ranked by protein coverage

Fig. 4

Functional classification of groups of proteins captured by the identified unique peptides for each GH family. Functional annotation of proteins captured by unique peptides was done with GhostKOALA. N/A = lacks annotation. A table of all EC numbers shown in the figure is presented in the Additional file 1: Table S4

Fig. 5

Classification of identified peptides by taxonomy using sequence information from the biogas microbiome MAGs. A Stacked bars show the classification of groups of proteins captured by the identified peptides, based on their taxonomical origins at the phylum level. Numbers in parenthesis are the total number of proteins in which the peptide is found. B Number of identified peptides required to cover all the proteins produced by members of a phylum for each GH family. NA, lacks annotation at the phylum level