Researchers from Scandinavia and the US have used proteomics technologies to create what they said is the most comprehensive map of proteins in the human gastrointestinal tract, providing a reference point for what may constitute a healthy human gut.
While the GI system contains a myriad of microorganisms that play an important role in digestion, little is known about the role they play in the defense against pathogens.
Metagenomics studies have shed some light on this process by identifying about 1,000 microbial species in the GI tract and providing information about the complement of genes found in the gut microbiome.
Such studies, however, provide only a glimpse of what may be happening in the gut because they provide no direct information about which genes are expressed or functioning.
In order to fill in the blanks, a team of researchers from Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, the Swedish University of Agricultural Sciences, and other sites performed a shotgun proteomics experiment to uncover more than 2,200 proteins, including new ones, in what they said is “the deepest coverage of a complex metaproteome to date,” according to an article published Oct. 30 in the online edition of The ISME Journal describing their research.
In it, the authors said that metagenomics studies have yielded important information about the human GI system. A 2006 study published in Science, for example, found that the GI microbiome has “significantly enriched capacities” for glycan, amino-acid, and xenobiotic metabolism, methanogenesis, and synthesis of vitamins and isoprenoids.
“This indirect evidence suggested that there are unique microbial functions carried out in the gut environment,” the authors of The ISME Journal article said.
Such studies, however, are limited because they leave gaps in information about gene expression and do not even tell researchers whether DNA found in the gut is from active cells or dormant, or even dead, cells, the authors said. But while a protein-based analysis would fill in such information holes, only a handful of such studies have been done, leading to the identification of only a “couple” of microbial proteins found in the gut, they added.
In their work, the authors aimed to develop a high-throughput, shotgun mass spectrometry-based approach to identify microbial proteins in a complex sample — in this case feces — and to “test the feasibility of using a non-matched metagenome dataset for protein identification,” they said.
The approach they created starts with the lysing of a cell mixture. The resulting proteins are digested and the peptides are separated by two-dimensional liquid chromatography. Finally, analysis is done on a mass spec.
Importantly, the approach does not call for the use of 2D gels, which allows it to be done in a high-throughput manner, Janet Jansson, the senior author of the study, told ProteoMonitor. The gel-free approach in combination with “very accurate mass readings” and relying on matches with sequence databases results in a workflow in which “you get thousands of protein identifications in just a matter of days,” she said.
A similar approach had been used only on simpler systems such as acid-mine drainage and sewage sludge water, but because their study looked at the much more complex fecal sample, Jansson and her colleagues needed to make some modifications.
For instance, the researchers used different kinds of sample processing steps in order to get the bacterial cells from the fecal samples, Jansson said. In addition, “compiling the databases was important so that we could get some accurate peptide identifications,” she said. “We also added some steps that weren’t done before to determine the relative abundances and the core proteomes by comparing these two individuals” from whom the samples were collected.
“If you look at the metaproteome, you really see what is functioning.”
By comparing their results against a database containing all predicted human proteins and the gut metagenome, the researchers detected 1,822 redundant and 1,534 non-redundant proteins. Of the non-redundant proteins, about one-third matched human proteins, while the rest matched proteins from the microbial metagenome sequence data.
They then matched their results against a database containing all of the sequences from the first database and sequences from “representatives of the normal gut microbiota,” including strains of Bacteroides and Bifidobacteria, and human pathogens and distracters not expected to be found in the human gut, such as environmental isolates. The rice genome was also included to help identify plant-related proteins.
That comparison resulted in the identifications of 2,911 redundant and 2,214 non-redundant proteins.
Comparing their findings with results from the metagenomic study published in the 2006 Science paper, Jansson and her colleagues found discrepancies in how the proteins were distributed among functional categories. The majority of proteins in the metaproteome, for example, are involved in translation, carbohydrate metabolism, or energy production. More proteins in the metaproteomes were also involved in post-translational modifications, protein folding, and turnover than proteins implicated in the metagenome.
Proteins associated with inorganic ion metabolism, cell wall and membrane biogenesis, cell division, and secondary metabolite biosynthesis were under-represented in the metaproteomes compared to the metagenome.
The overall picture from the metaproteome, Jansson said, provides a clearer and more accurate understanding of what is actually happening in the GI tract.
“This demonstrates that the metagenome is showing you all the potential for expression, but not all of those genes are expressed in the gut,” Jansson said. “If you look at the metaproteome, you really see what is functioning.”
The researchers then calculated normalized spectral abundance factors to find the relative abundance of the proteins they found. Common abundant human-derived digestive proteins such as elastase and chymotrypsin C were found to be the most abundant proteins.
The most abundant microbial proteins included those for expected processes such as enzymes involved in glycosis. Ribosomal proteins as well as DNA-binding, electron transfer flavoproteins and chaperonin GroEL/GroES were also “relatively” abundant, the authors said.
Unlike in the 2006 Science metagenomic study, in which key genes were implicated in methanogenesis, Jansson and her colleagues found “very few” proteins associated with methanogenesis in their study. Methanogenesis is an important process for H2 removal, but occurs only in about 30 to 50 percent of people in Western countries.
“In the remainder, H2 is consumed by sulfate reduction or reductive acetogenesis, and this seems to be the situation for the samples we studied here,” the authors said.
In an analysis of 205 unknown hypothetical proteins, they determined that most belong to novel protein families that are over-represented in genomes of gut microbes. Half of the 10 most abundant proteins belong to a novel protein family represented by the hypothetical protein CAC2564, and four of the top-10 belong to another novel protein family represented by the protein BF3045 from Bacteroides fragilis. The authors theorize that the two proteins are involved in amino acid metabolism.
While the majority of proteins identified were microbial, about 30 percent were human proteins, which Jansson said was an unexpected result. By identifying both microbial and human proteins in the same sample, the study provides a first glimpse into the complex microbial-host dialogue, she said.
Digestive enzymes and structural cell adhesion and cell-cell interaction proteins comprised the two largest groups of human proteins they found. A third large category comprised human innate immunity proteins, including antimicrobial peptides and scavenger receptor cysteine-rich proteins.
Jansson said she and her colleagues found one immunity protein, DMBT1, to be especially interesting. Predominantly expressed in epithelial cells and secreted into the lumen, DMBT1 is thought to have several beneficial functions including tumor suppression and anti-inflammatory effects. Their analysis suggests that the presence of complete, intact DMBT1 is indicative of a healthy gut.
“Although the human immune response is usually described in terms of response to infection, it is clear that innate immunity proteins are part of a normal gut environment, shaping the gut microflora to the desired shape,” they wrote.
In follow-up work, Jansson said that she and her team are using the same approach and looking at Crohn’s disease with an aim at developing a diagnostic tool.
“Here, this is just showing what’s normal, and the next step is to look at disease,” she said.
In the article, she and the other researchers said that though their study is the largest-scale study of the human gut microbial metaproteome to date, further work is needed to fully understand the how the human gut microbiota functions and interacts with the human host. One strategy to do so, they said, would be to increase the dynamic range of detection, though they add that “increasing the dynamic range is a clear challenge for all proteomic applications and this will be a pressing area for research and method development in the future.”
Jansson and her colleagues said that the goal of the study was to develop a method “to obtain a first large-scale” glimpse of the microbial community in the human gut.
While computational and experimental challenges remain, they said that their study shows the feasibility of their approach on a complex sample. Their study, they added, should be applicable to other complex environments such as marine and soil communities.