Background: Celiac disease (CeD) is increasingly common and affects around 1% of gluten-consuming populations1. Prior work has found that the composition of the duodenal microbiome is altered in individuals with untreated CeD (UCeD)2,3. However, analysis of duodenal microbial communities can be limited by contaminating nucleic acids and/or poor recovery of fastidious organisms4,5. To overcome these limitations, we developed a protocol using bioorthogonal non-canonical amino acid tagging (BONCAT)6 to identify translationally-active bacteria from duodenal aspirates.
Methods: Bacterial pellets from duodenal aspirates were anaerobically incubated with L-azidohomoalanine (L-AHA), a homologue of methionine, along with cycloheximide to inhibit eukaryotic translation. Following incubation, cells were stained with a Cy5 fluorophore conjugated to dibenzocyclooctyne (DBCO), which undergoes strain-promoted cycloaddition to the azide group in L-AHA. Using fluorescence-activated cell sorting, we sorted active (Cy5+) and inactive (Cy5-) bacteria from samples via comparison to a matched non-L-AHA incubated sample from the same patient. We performed shotgun metagenomics on paired active and inactive bacterial cells from patients with UCeD, treated CeD on gluten-free diet (TCeD), and non-CeD.
Results: Across all samples (active and inactive), bacterial community structure was distinct in UCeD and TCeD compared to non-CeD. However, analysis of active bacterial communities improved the resolution between non-CeD and CeD. The abundance of arginase was elevated in samples from UCeD and TCeD compared to non-CeD, suggesting that bacteria from patients with CeD have increased capacity for arginine degradation. In addition, there was a higher abundance of arginase genes in active versus inactive bacteria from most patients with both UCeD and TCeD.
Conclusions: BONCAT can be used to distinguish active and inactive bacterial populations from the human duodenum. Analysis of these populations suggests that degradation of arginine in the duodenum is increased in patients with CeD and remains elevated even with resolution of inflammation via GFD, particularly among translationally-active bacteria.